The moment you touch the new Framework 13 Pro laptop you know it's different to what's come before. Despite the bubblegum pink USB Type-C ports I've jammed in either side of the machine, this is a serious laptop for serious work. But it's also still instantly recognisable as one of the company's eminently upgradeable and repairable notebooks. This is the true beauty of Framework's new machine, a laptop that at once exists as the pinnacle of its hardware journey, but also steadfastly refuses to compromise on the core principles that has made it one of my favourite PC builders.

Though, it must be said, that refusal to compromise here comes at a hell of a cost because of the focus on ultra expensive LPCAMM2 memory.

This is a next-generation Framework laptop, but is still completely compatible with every Framework 13 device the company has shipped in the past five years or so. That is a hell of an achievement, meaning that you could have an ageing 13-inch laptop and pick and choose from the company's Marketplace store whichever parts of the new machine you wanted. If all you wanted was the beautiful new CNC aluminium chassis and input cover, go for it; if you just wanted the new touchscreen, be our guest.

But it's the full package here, with the Panther Lake-based Core Ultra Series 3 mainboard at its heart, that has made this hands down the best Framework machine you can buy or build. Every detail of the new laptop feels exquisite… bar one. And I know it's a familiar refrain from me now, but I wish there was a more premium display option. I've been spoiled by OLEDs and Retina displays, and if Framework wants to truly own its 'MacBook for Linux' tagline then that's the only place it falls down. And even then, only slightly.

Otherwise, I'm in love with what Panther Lake offers this device—almost silent operation, super-long battery life, and outstanding gaming performance—as well as how it feels, both in terms of that new chassis and this vastly improved keyboard and haptic touchpad. The Framework 13 Pro represents the idealistic young company all grown up, and it's a machine that's an absolute delight to use.

Framework 13 Pro features

From a pure performance and efficiency point of view, it's the Intel mainboard sitting inside the Framework 13 Pro which makes this such a powerful and usable laptop. The Core Ultra X7 358H chip baked into it comes with the magnificent Arc B390 integrated graphics core that has made the MSI Claw 8 EX AI+ such an outstanding gaming handheld, and that gives the 13 Pro some impressive gaming chops, too.

One other interesting thing to note about the new mainboard is its use of LPCAMM2 modules. Interesting… and expensive. When the design choices were made—with the goal of maintaining high memory bandwidth and throughput for the Panther Lake CPU without soldering on LPDDR5X and losing upgradeability or repairability—Framework would have known it was a more expensive option, but the memory crisis has made that exponentially worse.

There is also a level of fear to dropping in a new memory module that I've not experienced before. When you add in the LPCAMM2 module you need to line it up on a whole load of pins. Pins which, like motherboard socket pins, are very much incompatible with an errant thumb brushed across them. Too casual installation will kill a new mainboard real quick.

That new Intel SoC has been wrapped in a new cooler, which along with the efficiency gains of the Panther Lake architecture, allows for regularly silent operation. Combine that efficiency with the increased battery capacity of 74 Whr (compared to 61 Whr in previous versions of the machine) and you're getting stellar battery life out of the machine.

To cope with that increased battery girth Framework has redesigned the CNC machined aluminium bottom cover to have less of a taper to the front of it. The actual thickness of the frame hasn't really changed, but it's more uniform now. Which is no bad thing, and really does give it that MacBook/Razer Blade feel to the chassis. Impressively though, it is still backwards compatible with all the other Framework 13 laptops the company has ever released in that you could order the chassis on its own and drop your existing system into it without an issue.

Also getting the CNC aluminium treatment is the top cover for the rear of the display and the input cover housing the keyboard and the new haptic touchpad, both of which are excellent. The increased travel on the keyboard makes it a joy to type on, and the responsive trackpad is a real improvement over the original, and was designed especially for Apple folk coming across the divide to a PC laptop. My only concern is that multi-finger gestures too easily actuate a press, which is difficult to configure on the Fedora distro I'm using at the moment.

Finally, the Framework 13 Pro comes with a new touchscreen panel. It's a 120 Hz display, with a 2880 x 1920 native resolution that runs impressively bright (up to 700 nits) if you want it to. What it isn't, however, is a responsive OLED display. It's not a gaming panel, as CEO Nirav Patel explains to me when I speak to him on the show floor at Computex, but there is still a level of ghosting to the display which doesn't feel as premium as everything else about the device.

Framework 13 Pro value

The concept of 'value' for the Framework 13 Pro is a tricky one. On the one hand you have the fact that the best version of this machine comes with the Intel Core Ultra Series 3 processors, and they demand the fastest LPDDR5X memory you can give them. And, to ensure Framework sticks to its upgrade/repair ethos, it is using LPCAMM2 modules to deliver that fast LPDDR5X memory.

These are not common, and they are not cheap. There are few suppliers, and Framework recently admitted that its current supplier for these Micron modules has increased the cost by "more than double". You only have to look at the configurator on the Framework website for the 13 Pro to see what that means. A 16 GB LPCAMM2 module costs $239, while the 32 GB version adds another $800 onto the cost of the laptop. That makes the cost of the Framework 13 Pro rather prohibitive for most mere mortals.

Even a 16 GB 13 Pro config will run you well over $2,000. And even if you switch to AMD, the HX 370 Framework 13 Pro configs of the same spec costs just shy of $2,500.

The other side of this, however, is that Framework has been walking the walk on long-term upgradeability for its devices for many years over many different generations of PC hardware from different manufacturers. That means the initial premium outlay for a Framework system can be offset by the fact that you may not have to buy another whole laptop again for a very, very long time given the level of granular upgrades on offer.

I will grant you that this is a very high bar to get over, however. When the cost of an initial Framework 13 Pro purchase is this high you're going to have to keep hold of this thing, upgrading piece-by-piece for many generations, to even get close to breaking even with someone who just bought a new machine every time, and hang the ecological consequences.

Framework 13 Pro sound

When it comes to the actual audio experience, Framework has added new side-firing speakers to the design, rather than the original down-firing ones. They deliver impressive clarity and will go pretty loud, too, but they are still just laptop speakers. Basically, do not expect real room-filling audio with thumping bass.

But it's the other sound production side of the Framework 13 Pro that has me truly impressed, and that's the fan noise. It's long been an issue for me and Framework laptops; there's no way to control the fan levels unless you hack in some third-party tools made by the community, and they noise has often trended far too far to the unpleasant for my liking, especially when under serious load.

That was a concern for me as someone who wanted to use the B390 iGPU at the heart of the Core Ultra X7 CPU inside this machine for gaming. I shouldn't have been worried, however, as even when I'm pushing the Arc GPU the system is tuned to rein it back in a little (but not too much) and keep it running at a genuinely quiet level. There is fan noise, but it's slight at worst.

For most of the time I've been using the Framework 13 Pro outside of gaming, though, it's been all but silent. And often actually silent.

Framework 13 Pro performance

When I first started benchmarking the Framework 13 Pro, with its Intel Core Ultra X7 358H chip, I'd come off the back of testing the MSI Claw 8 EX AI+ with the very similar Intel Arc G3 Extreme processor. And there was a slight level of disappointment at seeing the performance not quite hitting those levels or the levels of our initial Panther Lake performance tests when the architecture first came out.

But I've already mentioned just how quiet a device this 13 Pro machine is, and that's as true when doing CPU rendering as it is when I'm pushing it with Cyberpunk 2077. On a machine such as this I'll take a few missing frames per second in exchange for an actual experience that doesn't make me want to stab myself in the ears to get away from the sound.'

You can see in the CPU benchmarks that the chip is bouncing off 100 °C and thermally throttling, and it's the cooling limitations necessary to maintain the chilled experience which are slightly holding back the gaming frame rates, too.

Even if we are talking about a slight tempering of the performance of the Arc B390 iGPU compared with other devices sporting Intel's top integrated graphics, it's still the most powerful that you will find and places this machine head and shoulders above the best that AMD's Ryzen AI 9 HX 370 can muster.

And I was already pretty impressed with that.

Framework 13 Pro aesthetics

With its graphite black, CNC milled aluminium chassis the Framework 13 Pro is the most effortlessly stylish laptop the company has ever produced. It feels like the culmination of everything Framework has been working towards for these six years. The slightly chunkier styling, necessitated by the bigger batter sitting underneath the touchpad at the front of the device, makes it feel far more solid than before.

And actually makes it feel far more like the MacBook it's trying to ape, and the Razer Blade 14 that it's potentially going to replace as my daily driver.

It actually is more solid, too, as you can pick it up by the corner and the extra structural rigidity afforded by that new aluminium frame means there is zero give. The same is true of the keyboard as well which betrays no movement bar the extra travel of the keys Framework has deliberately introduced into this version.

But then you've also got all the other Framework touches that has made its 13-inch laptop such a pleasing device for years. You have your choice of coloured and translucent bezels as well as a plethora of colours for the USB Type-C input modules. For this layout I've gone for a pair of bubblegum pink ports and a translucent grey bezel for that extra techie feel. But that's my preference and you can make your own choices at the point of purchase yourself.

Framework 13 Pro software

There is no bloat, because there doesn't have to be any software included at all.

If you just want to stick a Linux distro onto your device the SSD will be shipped blank, likewise if you want to use your own copy of Windows rather than spend $200 on a needless Windows Pro license.

There is a driver bundle for Windows you can download from the Framework support site, and you will need Wi-Fi drivers in hand when you install Windows ('cos that's still a thing), but Linux support is largely there out of the box when you install your distribution of choice.

The slight downside of this, as I noted before, is that there is no system control software. That means there is no granular control over power options beyond the OS options, and there are no fan controls, either. The fan control issue isn't a thing for Panther Lake, however, thanks to its quiet operation straight off the bat, but I wish I could control the SoC's TDP as easily as you do on a gaming handheld, but that's a relatively minor gripe in a world that is often replete with too many different bits of control software fighting each other.

Framework 13 Pro battery life

Of course this is not a gaming laptop, but the level of gaming performance you can get out of this ostensibly productivity laptop is excellent. Development by day, gaming by night. Or on the go, because the battery life of the Framework 13 Pro is also very, very good.

Stick it on balanced mode in Windows and boot up a game and you'll find the battery drains in a little over two hours, but the latest Arc drivers include the same Endurance mode I loved so well in the MSI Claw 8 EX AI+. It doesn't quite get you the same five hours of gaming that handheld delivers, but you're still not getting far off a doubling of the gaming uptime of the standard mode.

That does try and stick you to 30 fps as a rule, but when you're trying to keep gaming as long as possible when you're on the move—or trying to make sure that when you're done playing you have some power left for some actual work—then it's a worthy compromise.

That, however, is a Windows thing, requiring the latest Intel Arc drivers and isn't necessarily something you can replicate within a Linux environment. Almost all of Framework's impressive 13 Pro battery life numbers are recorded from a Windows installation, with only its seven days of standby measured with Ubuntu.

That doesn't mean it's bad in Linux, not by a long shot. I'm using a Fedora Cosmic Atomic distribution at the moment, and have been running it all day long today, without it going into standby once, playing music, downloading games, and having video meetings, as well as writing this review, and I've only gone through half the battery capacity in six hours. It's all-day battery life for sure, though I think Panther Lake's optimisations are probably going to be felt more keenly in a Windows environment.

When most people are looking for a gaming headset they want it to have a few things: work with their platform of choice, sound good, decent mic, and a reasonable battery life. Other features like noise cancelling are nice but unnecessary, as well as adding to the price of the unit. The LucidSound LS500 does away with such frivolities and focuses in on what you really need in a headset, delivering a surprisingly good pair of cans for under $70 USD.

Taking the headset out of the box the first thing I noticed is how light it is. The LucidSound LS500 boasts long haul battery life of up to 70 hours, so I was expecting a heavier fit. Instead it is a lightweight and comfortable headset. This is thanks to the plastic construction, which feels a little cheap but makes up for it in featherweight comfort. It even includes glasses slots to make long wear for four-eyed folks like me.

It’s not as comfortable as Turtle Beach’s solution but does a far better job than the likes of top dog Razer BlackShark’s for glasses comfort. Despite this, they are living up to their battery life quotas as I have yet to charge this headset after many tens-of-hours of testing.

This lightweight plastic construction delivers a headset that weighs only around 260 grams. The mesh construction headband also does well for long hours of gaming. If it’s not my glasses tiring me it’s pressure on the head, and I can say this headset stays comfortable for many hours-long gaming sessions.

Connectivity will depend on the unit you have, either Xbox or PlayStation, but both come with a handy dongle that can also be used on your PC. The dongle is quite large, but has USB Type-C and Type A connectors which is a nice touch for versatility. The 2.4 GHz Connection can be activated by holding down the ear cuff button for a few seconds, or you can swap to Bluetooth mode using the power button. Unfortunately, you can’t do both at once, but at this price I wasn’t expecting anything that fancy.

It is worth noting that the chat boom style mic doesn’t appear to work in Bluetooth mode, and only responded when I was gaming with the dongle connection. In game it works just fine, with mic monitoring available by the other ear cuff button—though it’s never worked for me. Still, for a headset this cheap, the mic has fine quality pickup with pretty good voice isolation, and I was having no problem clearly shouting out enemies and mining goods in our Deep Rock Galactic adventures.

But how do they sound? Really quite great, especially for the price. I spent hours just riding my horse around in Ghost of Yotei on PlayStation 5 just enjoying the sounds of nature and the breeze through fields. When coming to campsites, I’d be able to ride around them and the directional sound always pointed me in the right direction. In Hades 2 on PC I was able to turn on the vow or rivals and locate Hecate’s true form by the direction of her voice.

When listening to music, sure other headphones would do a better job, but these really aren’t bad. There’s a surprisingly good breadth of sound, with treble and bass both coming through clear and well tuned. Though the latter is a little lacking especially if you’re used to bass boosted cans. Still, I doubt you’d get much better for a regular, non-gaming headset that cost around the same amount of money. They even look reasonably stealthy with the boom style mic in the upright position. It’s certainly a gamer headset but doesn’t go too hard on the gamer aesthetics so you’re not embarrassed to wear it out in public.

For under $70 USD, the LucidSound LS500 provide a really well rounded headset that has focussed on the most important qualities to gamers. Having such great audio quality for this price really is a huge win. All the sacrifices that have been made to get this headset to the price point are really worthwhile ones, leaving just what you need in a quality gaming headset. Great audio, a decent mic, and the ability to connect to whatever you need.

From the archives: This story originally ran in PC Gamer #321 (UK, Sep. 2018). The review appears as originally written, with only minor changes in formatting and presentation.

It’s a damn crime that no game ripped off Red Faction: Guerrilla’s Geo-mod destruction feature. You take a hammer and you use it to smash out a panel of a wall and break through a pillar. Smash out a few more walls and maybe the building creaks. A few more whacks of the hammer later and the whole thing comes down, perhaps with you inside. And hey, maybe it knocks through another building on the way down. It’s still amazing to watch.

Guerrilla does this one thing really well. It’s an otherwise uninspired open world game, with okay driving, unimpressive shooting and a boring story. Hell, Mars doesn’t even look that nice, because it’s Mars. But knocking down buildings in an open world is more than enough to carry this game.

You’ll crash a jeep through a base. You’ll detonate mines and blow up a bridge, which will then collapse on top of an enemy settlement. You’ll melt the beams of a tower and watch it fall over. This destruction felt great in 2009, and it still feels good now.

In this Re-Mars-tered edition, which is free for owners of the Steam version, THQ Nordic touts reworked textures among other improvements, along with better shadows, lighting, a shader and postprocessing rework, and native 4K support.

For a freebie, it seems pretty good after four hours or so. I can’t say I notice vast visual improvements, but environmental textures like rocks and the ground look a little sharper. The lighting is nice, too, but it’s not a vast upgrade.

To anyone picking this up for the first time, I recommend focusing on the campaign. That’s where the biggest destructive opportunities are, like a huge bridge and a massive tower, which take meticulous use of the object-melting nano rifle and rocket launcher to bring down.

You’ll wait a while to unlock all the really good toys, but the singleplayer will give you plenty to blow up along the way, as well as a dull story about miners rebelling against an army that’s a bit like a boring version of Firefly. The big destructive opportunities in this campaign need to be seen, though—I still talk about them with friends years later.

I’m not in the mood for finishing the campaign these days after doing it twice before, but that’s where the game’s Wrecking Crew mode comes in. It’s basically a score attack that gives you a quick dose of the game’s destructive physics in a variety of settings.

You pick a loadout, set the parameters, like time limits and how easily the buildings fall, then knock them down as quickly as possible. It makes a great pass-the-pad party game, and there’s a challenge mode with leaderboards, too.

No imitators

It’s weird that no one else made a game like Guerrilla after its release. Destruction features have shown up in everything from Battlefield to Minecraft to Just Cause, but Volition’s game still offers something that no other game does. It’s the idea of blowing each building up like it’s a puzzle, efficiently using your arsenal so it collapses in the most satisfying way possible.

It might’ve been Red Faction’s sparse Mars setting that made Guerrilla work in the first place. It’s noticeable that all the best buildings are far away from one another, and it probably wouldn’t have been possible for Volition to replicate the Geo-mod system in the vast cities of Saints Row. But that is the type of open world game I’ve always wanted to play, where you can manipulate the environment.

We’re instead in an age of open world games packed with busywork and towers to climb. I like those, too, but Guerrilla shows open worlds and destructible buildings are a perfect match.

Even old THQ didn’t seem to know why Guerrilla was good. Volition followed this up with Armageddon in 2011, a bad sequel that took us away from big open setpieces to the tedious, more linear underground. Still, if THQ Nordic is bringing Darksiders back for another shot, perhaps Red Faction will get its chance again, too. I just wish its influence, and its collapsing buildings, had carried a little further.

After you spend above a certain amount of money on your gaming keyboard, you're not really paying for quality, specs, or competitive edge anymore. I felt this recently while testing out Melgeek's Centauri 80. In my approximation, any more than $200 and you move into bragging rights/centrepiece territory, and by that logic, I reckon the Asus ROG Azoth Extreme Edition 20 is worth about three centrepieces.

It's an expensive old beast, and $100 more than the already very pricey Asus ROG Azoth Extreme, which we reviewed back in 2024. But it's hard to say that Asus hasn't thrown absolutely everything at this new keyboard to celebrate the 20-year anniversary of ROG; you're getting an 8,000 Hz polling, up to 1,600 hours of battery life, a carbon-fibre positioning plate, as well as multi-layer dampening, plus a whole host of extra accessories and extras in the box.

Oh, and did I mention the 24K-gold-plated commemorative nameplate? I probably should have mentioned the 24K-gold-plated commemorative nameplate.

The Azoth Extreme is a very well-built bit of tech. It's super hefty, but it's heavy for a reason. It's sturdy, and though I wouldn't much like to drop a $600 gaming keyboard, I feel like I could do so with relatively minor problems here. The OLED screen up at the top right looks lovely, and the controls next to it are easy to use and quite handy. You can adjust volume, change brightness, and even set up custom controls. It's not quite as useful as the likes of the Corsair Galleon 100 SD and its built-in Stream Deck, but it will likely be enough for most.

Thanks to that weight, this TKL keyboard isn't exactly the type of thing you'd throw in a backpack. That said, a heavy keyboard is not nearly as much of a deal breaker as a heavy mouse or a heavy headset. In fact, the only thing I'd watch out for is the long-term wear on a particularly light mouse pad; 1.5 kg is a lot to put on any pad, and that's not including the weight of your hands on the wrist rest either.

The Azoth Extreme Edition 20 splits its keycap designs between standard black ones with a translucent base for the letter and number keys, then fully translucent caps for the F keys, space bar, and those on the side. Aesthetically, it works well and allows the per-key RGB to really shine. Initially, that RGB comes in all-gold (or, I suppose, with it being RGB, closer to a yellowed white). It makes the gold accenting on the side of the keyboard and bottom of the palm rest really pop. Despite gold tending to be a very gauche accent choice, it actually looks great in-person.

The two types of keycaps don't just add to the aesthetic; they also feel different on the fingertips, which is a decent way of keeping your fingers in check when you're not actively looking at the keyboard. The full translucent keys are very smooth to the touch, feeling similar to the ROG Harpe II Extreme Edition gaming mouse, while the other style of keycaps have more of a matte feel. Both keycap designs are pretty in their own right, and the difference in feel between the two isn't distracting either.

As far as typing is concerned, I'd say the Azoth Extreme Edition 20's actuation is on the light side, but not so much that I found myself misclicking. It's near perfect for accuracy, speed, and ease of use, in my book. The typing sound is light and clicky, with those dampening layers taking out any thock you might expect from the weight of shoving keys down. It's almost like Asus has scooped the bass and mids from it and just left the treble on top.

Typing feel, as a result, is lovely. Keys press easily with a light, satisfying sound. It's easy to find your place on the keyboard by feel too due to the mixture of keycaps and solid spacing all around. I do miss the extra keys of a full-fledged keyboard, but with some smart custom controls, I get by just fine.

My first gaming test with the Azoth Extreme Edition 20 was to take it into a Counter-Strike 2 match, and I can happily report that me losing had nothing to do with the keyboard. It's super snappy and responsive, with its light actuation keys making quick reactions much easier. It's also just plain effortless to use, with little fatigue after hours of gaming. As I learned when I tested the Melgeek Real 81, that's a really important facet of using any gaming keyboard.

You're getting an optional 8,000 Hz polling here (though worse battery life as a result), but you're still getting mechanical switches, which is a surprise. Them being hot swappable certainly helps, but one would hope for analogue or optical switches, for faster clicks, better battery efficiency, and rapid trigger. Still, it's super fast anyway, so you may not notice that downgrade from some other high-spec 2026 gaming keyboards.

Overall, I'd say the Extreme Edition 20 is a stronger package than the 2024 Azoth Extreme, coming with the poll rate booster, plus tons of extra bits, a palm rest, and an upgraded battery life. The NX mechanical switches feel great under the fingers, but you are paying $100 for that upgrade.

One problem I do have with the Azoth Extreme Edition 20 is a pretty common one, as far as keyboards go. The keycaps that don't have translucent tops have white lettering on black, which can be quite tough to make out in the dark. Some keyboard manufacturers get around this by making lettering translucent and shining RGB through it (like the Logitech G915 X Lightspeed), but that makes them harder to see under natural light. This means when there's little light on the keyboard, I have to make out keys through feel.

Unfortunately, Asus's software continues to be a problem. It worked fine when I first set it up but, it suddenly just stopped working when I tried to swap the OLED screen animation. The web app requires an extra bit of software to actually run, and it soon stopped as it said I didn't have the software, despite me installing it. A fresh install did not fix this problem, either. In Armoury Crate, it simply refuses to recognise the keyboard, even with multiple restarts and a fresh batch of software.

Software problems aren't unique, but the Azoth Extreme 20 is a glorious keyboard with golden highlights and a breadth of extra little goodies included in the box. It feels like a limousine coming to pick you up, but inside it turns out to be a plain old Honda Accord. There's nothing wrong with an Accord, but Asus is pulling out the stops for this keyboard, and it simply can't make it down the runway without a hitch.

Stating that the Azoth Extreme Edition 20 is expensive feels so obvious it's almost not worth saying. That might be kinda the point here: it's a limited edition bit of tech that comes with a thank you card, extra switches, a switch and keycap puller, a polling rate booster, the aforementioned nameplate, a cleaning cloth, and all manner of little extras. It is extravagant and meant to feel that way.

As such, comparing it to a standard gaming keyboard is a bit hard, as you don't walk into a Walmart and think 'I wonder if that's worth it'. It's more of a piece of hardware you search out, and if you're prepared to part with all of that money, you are getting a truly lovely bit of tech. I just couldn't ever see myself being prepared to spend that much on it.

I submit that there are two types of videogame remakes. The first type goes all in on transformation, taking big swings with its source material to the point that the remake can be embraced as a wholly new take. The most successful version of this approach is Capcom's Resident Evil remakes—horror shooters so beloved and distinct that they comfortably coexist with their originals.

Halo: Campaign Evolved isn't that.

What we have here is the other type of remake—the kind that goes all in on modernization, offering perceived upgrades to an otherwise unchanged foundation. Halo Studios has no discernably new take on Bungie's landmark FPS, other than a desire to make Combat Evolved look and play more like its later Halos. The "faithful" treatment sets an admirable goal, but that also means Campaign Evolved positions itself as a replacement for, or at least in competition with, the original.

That's a hurdle that was always gonna be tough to clear, because OG Halo is still a fantastic game 25 years later, and among my all-time favorite shooters. It's also readily playable with modern conveniences through The Master Chief Collection (another Halo Studios joint). I completed a playthrough of the original after running through the remake, which was helpful for spotting differences, but honestly, I'd take the weakest excuse to play more Halo.

The high value and quality of the MCC makes Campaign Evolved the superfluous add-on from the jump—it has fewer features, fewer campaigns, and zero competitive multiplayer.

New ride

What Campaign Evolved does have going for it, primarily, is novelty. It's neat to scour battlegrounds I've explored at least a dozen times reimagined by people who care, and to kick back to expanded and expensive-looking cutscenes that rival Halo 2 Anniversary's work. I especially admire how Halo Studios carefully applies hindsight to Halo: CE's weapon sandbox, allowing the occasional SMG, needle rifle, or energy sword to appear earlier in the series than expected without upstaging vanilla loadouts.

It's also impressive to look under the hood and consider what Halo Studios has pulled off here. With the help of underlying Bungie-era code, they've made Halo feel like Halo in an Unreal Engine 5 wrapper. If they hadn't done that right, the whole thing would've been dead on arrival, but the grenades do bounce correctly and Master Chief is appropriately floaty. Campaign Evolved is the best-feeling Halo that this team has ever assembled, and by extension, one of the slickest shooters of this scale in recent history.

Broadly speaking, it's a decent audition for future Halo projects (more remakes, right?) that I expect to go over smoothly with many Halo fans. But I can't help but also see everything that's worse about this thing—stuff that you could write off as nitpicks, but I see as evidence of this team's boring, more ordinary vision for Halo.

Master grief

Indeed, Campaign Evolved has the guts of Bungie, but that hasn't stopped 343 Halo Studios' historically flawed taste in Halo from gumming up the works, such as:

• A sprint button that defangs Flood encounters and makes vehicles less dangerous
• An assault rifle with such a long effective range and ubiquitous ammo that you're rarely pushed to use anything else
• A Warthog that's so convenient to drive and difficult to flip that the campaign's famous "Warthog run" loses all of its tension
• Waypoints that tell you exactly where to go at all times, despite level design already doing that naturally

They're bewildering choices made extra annoying by the remake's measured hand in other areas. Halo Studios clearly believes some old design principles are sacred—there's no mantling or grapple hooks, and only a tasteful sampling of later weapons. So to commit to these consequential revisions tells me the studio is not just updating a 2001 videogame for a younger palate, but truly believes that Halo is better this way. If so, they're wrong.

I know complaining about features that have become standard practice of an FPS sounds silly, but I'm unconvinced that Halo is improved when you can always run away from problems, or when Master Chief's mechanical steed controls like a Honda Civic, or when exploration is optional.

Yes, you're kinda slow in the original, but it's an appropriate speed for its mostly indoor arenas, and outside, Chief's jog ensures vehicles are important. Sure, the old Warthog flips a lot—that's why it's fun to learn its quirks and eventually get good at powersliding its massive body around corners. Yes, it's possible to get lost in the old game, but reading environmental cues can also be a rewarding part of play, and the remake doesn't even try to maintain that (you can turn off waypoints in HUD settings, but you can't disable them at a lobby level).

To enjoy Halo's classic qualities and miss them within this remake is not to succumb to nostalgia. It's to recognize the value of an FPS that does its own thing—that asks something slightly abnormal of its players for a real reason. I want to believe the stewards of Halo also think that, but my faith is shaken when modern trends are applied to a 25-year-old game as if they're an inevitability of an imaginary design rubric. I don't play Doom and wish for a realistic reload button. I don't play Mario and balk at a missing grapple hook. I don't play Halo 1 and expect to sprint.

Postcardification

I'm far less scandalized by, but still mixed on, Campaign Evolve's big art pass. This is an undeniably attractive game. I caught myself many times admiring Halo Studios' reassembled vistas, staring at a downright ridiculous ocean wave effect, or savoring cutscenes that highlighted Master Chief's Mark V armor in fully realized detail.

The results are as impressive as they are expected. I'm confident that this is exactly what Halo Studios was going for—a photoreal reimagining utilizing every bell and whistle that UE5 affords. But as time with Epic's latest tech has revealed, this is an engine that seems to excel at (or maybe insists upon) a certain "generically real" look.

Where I used to gaze out at Alpha Halo's horizons and see a misty, hauntingly beautiful facsimile of nature floating in space, I see in its place picturesque compositions of mountains, trees, and water that belong on a gift shop postcard. This remake makes Halo, the place, less alien.

Playing both versions back-to-back, I took more notice of background environments than I usually would. I like how both games look in all of these instances, but I tended to prefer the original for its readability, atmosphere, and color choices.

Define 'better'

Sometimes, Campaign Evolved would knock my socks off with a view that wasn't just beautiful, but dramatic and communicative in ways the original wasn't capable of, like the jungles of 343 Guilty Spark:

And other times (honestly, more of the time), UE5's raytraced light bounces would produce annoyingly dark rooms with barely anything to appreciate, like in Silent Cartographer:

Rooms like that gave me a renewed appreciation for hand-crafted, baked lighting placed exactly where a developer wanted it. Every room in Halo: CE was the work of fine brush, but Campaign Evolved gives the impression of broader stroke simulations shouldering the VFX load, often with diminished results.

I had the same reaction to the new rendition of Forerunner facilities. In keeping with Halo Studios' past games, the remake's interpretation of Forerunner architecture is of highly reflective steel, made somewhat brushed and weathered only after fan feedback. Not only are these materials less distinct and alien than the original's rough, intriguing marriage of brutalist concrete and metal, but those fancy reflections also turn indoor gunfights into overstimulating particle fests. It never gets so bad as to be broken, but it was an unwelcome obstacle in fights. It's as if readability wasn't the priority for Halo Studios that it clearly was for Bungie.

Speaking of the B word, the close timing of Campaign Evolved and Bungie's Marathon reboot is a useful contrast. One of my favorite things about Bungie's extraction shooter is how its gorgeous environment design zags totally away from photoreal, betting all of its chips on readability, color, and dramatic weather effects that masterfully sell its alien world. The entire visual identity of Marathon ignores the presumption of "next-gen" graphics becoming a slurry of tech demo trees and mountains, and flies in the face of the narrative that Campaign Evolved's new look represents "what Halo would've looked like if Bungie was able."

I also ding this remake for artistic changes that make Halo less fun, like how they messed up plasma grenades. In the remake, the sticky nade has lost its gaseous visuals and dampened its signature hissing, important signs for noticing when one is about to go off. I know because this is the most I've ever died from plasmas in a Halo game, and it was usually because I was standing right next to one without realizing it.

Then there are updated gun sounds that I'm comfortable calling downgrades. The beam rifle has lost its iconic "shutter click", substituted with commonplace laser noises. Halo Studios, post-feedback, repurposed the assault rifle's classic cacophonous scream, but lowered its presence in the audio mix.

And the sniper rifle. Hoo boy, what have they done to my sweet child?

Extra time

The best thing about Campaign Evolved is that you still, ultimately, get to play one of the best shooters ever made. That is points in its favor, though I doubt its proposed value to anyone who's already acquainted.

Since I've seen confusion online and even among friends excited for this remake, it's worth laying out exactly what you do and don't get for $50:

✅The Halo: CE campaign, remade
✅3 bonus missions set before Halo: CE
✅4-player co-op (the original only supports 2)
❌No PvP multiplayer whatsoever
❌No theater mode

It's a sparse package befitting its literal subtitle. While 4-player co-op is a big add, the absence of proper PvP is a serious blemish that ignores a huge reason people love these games, and will personally limit its lasting appeal. I've been denied an Unreal 5 remake of Blood Gulch's skybox, and that must be some sort of crime.

And it can't be ignored that The Master Chief Collection is $40 (often much less on sale) for much, much more:

✅Six Halo campaigns, two with Anniversary updates
✅Full multiplayer support for all games with crossplay
✅Forge, Theater, File Share
✅Mod Support with Steam Workshop
❌No bonus missions

Slightly dulling the blow are those three bonus story missions, though they didn't turn out as fun as I hoped. The trio of lengthy assignments take place on a single Covenant ship, resplendent in shiny purple hallways with little variation beyond a middle chapter that goes outdoors.

I found Master Chief and Sgt. Johnson's quipping in these episodes out of step with their dynamic in the main series, but worse was a noticeable drop off in the quality of encounters. The close quarters and reliance on door spawns did a poor job of hiding the seams between enemy waves, and battles lingered longer than was fun.

The never-ending fights chafe against wonky difficulty tuning—on Heroic (my default), there's not enough ammo or cover to reasonably deal with bullet spongey brutes solo. It takes a lot to push me over the edge in Halo, but when I finally did switch to Normal, suddenly the challenge was in line with the base game. I suspect a balancing patch is in order.

The highlight of the optional interludes was weapon variety—this is where Halo Studios' lets you cut loose with the later-game guns like the battle rifle, Reach's needler rifle, and the brute plasma rifle. There's also one area in mission one that's basically an indoor skate park for Ghosts—a super cool idea for a Halo encounter.

Campaign Evolved ran very well for me at 1440p (90fps average without upscaling), but honestly, that's not an illuminating observation when you play on a RTX 5090, Ryzen 7 9800X3D, and 64GB RAM. I kept everything on Ultra and didn't babysit a framerate counter. I will say that, as the very best case scenario for hardware, I didn't like how fuzzy DLSS can still make bright light sources look. It was distracting enough that I switched to native rendering and took the frame hit. Even though the system requirements were no impediment to me in the review, I wonder if my friends on middling hardware will want to pay to play a version of Halo that runs worse than the one they already own.

That scenario gets to the heart of it: the next time I want to play the first Halo, I don't have many reasons to reinstall this 70GB remake. It was cool to see another interpretation of a game special to me, but the glitzy coat of paint primarily highlights how right Bungie got it in 2001, and how its modern stewards are nudging Halo in an unremarkable direction.

The Arctic Freezer 36-S lands at the right time. Prices are rising for PC builds but you can claw some money back with this cooler. It costs just $22/£19. A small price to pay for ample cooling for your processor. And it is, genuinely, good.

Arctic has plenty of accolades to its name when it comes to cooling. One of its more popular designs, the Arctic Freezer 36, is our pick for the best budget air cooler. With two fans, an effective heatsink design, and a great price tag, it's all anyone needs to chill most of today's popular gaming chips. But could you make do with less? That's what the Freezer 36-S puts to the test.

A few cuts have been made to the Freezer 36-S compared to the Freezer 36, including trimming down to just one fan. But it still has a lot going for it, especially when it comes to performance, as you can see in the charts below. The Freezer 36-S performs really well considering its size and price tag. Though price is a sticking point: the Freezer 36-S is going up against the best and when you can get the Freezer 36 for very nearly the same cost, it's tough to argue for the slightly cheaper model.

Features

The Freezer 36-S is a fairly slim cooler at 126 x 88 mm. It's a pretty traditional-looking thing, with a similar height to the Freezer 36. That makes it a fit for standard ATX and Micro-ATX rather than small form factor cases.

There's only one fan on the Freezer 36-S, a numerical loss to the two on the Freezer 36. Though it's a P12 Pro fan, an improvement on the twin P12 PWM PST included with the latter. This P12 Pro fan spins up to 3,000 RPM and comes with impressive ratings for both airflow (77 cfm) and static pressure (6.9 mmH2O), the latter being important once the fan is stuck to the side of a heatsink.

The P12 Pro is our current pick for the best budget PC fan and performs exceptionally well in testing throughout its large RPM range. I do have one recommendation, and that's to adjust the fan curve in the BIOS. These fans are incredibly loud at top speed, but they don't need to run this fast to be effective. They're very capable and much quieter at slower speeds.

One accessory bundled with every Freezer 36-S is a contact frame. A contact frame is just a piece of metal or, in this case, plastic, which sits around and atop of a processor instead of the standard ILM. It's important for Intel processors from the 13th/14th Gen as they are likely to bend due to the stresses applied to the chip from the ILM and cooler. To avoid this, a contact frame is preferred, and Arctic bundling one with the Freezer 36-S takes the stress out of it. That said, the metal one included with the Freezer 36 feels more robust.

Arctic throws in a small 0.8 gram tube of MX-7, the company's latest thermal paste. It's good stuff, and there's enough here for initial installation and at least another application, providing you apply a thin line and don't go overboard.

Installation

Installation is fairly straightforward with the Freezer 36, if only a little tricky depending on whether you're using an AMD or Intel processor.

The good news is it couldn't get much easier on AM5/AM4 systems. Fit a couple of brackets, apply some thermal paste, and tighten the cooler into place using two screws. Sorted.

Intel systems, on the other hand, require the removal of the built-in ILM from the motherboard and replacing it with the contact frame. That's easily done ahead of installing the motherboard inside a case, less so if not. If only humans had one more arm, it'd be a lot easier to hold everything in place while screwing the frame down. There is an added risk of the CPU falling out or breaking a pin in the socket with the contact frame, but just take it slow and it should be fine. Having spoken to Arctic about this earlier in the year, they're convinced that the contact frame is worth it in the long-run.

The Freezer 36 comes with two metal clips for holding the P12 Pro fan in place. These clips attach to the cooler itself. Unlike most other coolers I've used, which do the opposite. Install the clips first, then the cooler, than clip the fan on. This way, nothing ever gets in the way and you don't have to fiddle clips in hard to reach places. Nice design.

Thermals

Thermals

Evaluating cooler effectiveness using Core i7 14700K

Baldur's Gate 3 @ 1.275 V | Normalised to 20 °CX274 @ 1.275 V | Normalised to 20 °CPrime95 @ 1.275 V | Normalised to 20 °C | ~170 WPrime95 @ 1.325 V | Normalised to 20 °C | ~220 WAmbient temperature

Max °C

Avg °C

Arctic Freezer 36-S A-RGB

54

51

Arctic Freezer 36

54

51

Be Quiet! Dark Rock 6

54

50

Be Quiet! Dark Rock 5

57

53

Be Quiet! Dark Rock Pro 6

54

48

Thermalright Peerless Assassin 120 SE

52

48

015304560

Baldur's Gate 3 @ 1.275 V | Normalised to 20 °C Data

Product	Value
Arctic Freezer 36-S A-RGB	54 Max °C, 51 Avg °C
Arctic Freezer 36	54 Max °C, 51 Avg °C
Be Quiet! Dark Rock 6	54 Max °C, 50 Avg °C
Be Quiet! Dark Rock 5	57 Max °C, 53 Avg °C
Be Quiet! Dark Rock Pro 6	54 Max °C, 48 Avg °C
Thermalright Peerless Assassin 120 SE	52 Max °C, 48 Avg °C

Max °C

Avg °C

Arctic Freezer 36-S A-RGB

64

56

Arctic Freezer 36

63

55

Be Quiet! Dark Rock 6

63

54

Be Quiet! Dark Rock 5

68

58

Be Quiet! Dark Rock Pro 6

63

63

Thermalright Peerless Assassin 120 SE

60

52

020406080

X274 @ 1.275 V | Normalised to 20 °C Data

Product	Value
Arctic Freezer 36-S A-RGB	64 Max °C, 56 Avg °C
Arctic Freezer 36	63 Max °C, 55 Avg °C
Be Quiet! Dark Rock 6	63 Max °C, 54 Avg °C
Be Quiet! Dark Rock 5	68 Max °C, 58 Avg °C
Be Quiet! Dark Rock Pro 6	63 Max °C, 63 Avg °C
Thermalright Peerless Assassin 120 SE	60 Max °C, 52 Avg °C

Max °C

Avg °C

Arctic Freezer 36-S A-RGB

66

62

Arctic Freezer 36

64

60

Be Quiet! Dark Rock 6

63

59

Be Quiet! Dark Rock 5

70

67

Be Quiet! Dark Rock Pro 6

63

61

Thermalright Peerless Assassin 120 SE

64

58

020406080

Prime95 @ 1.275 V | Normalised to 20 °C | ~170 W Data

Product	Value
Arctic Freezer 36-S A-RGB	66 Max °C, 62 Avg °C
Arctic Freezer 36	64 Max °C, 60 Avg °C
Be Quiet! Dark Rock 6	63 Max °C, 59 Avg °C
Be Quiet! Dark Rock 5	70 Max °C, 67 Avg °C
Be Quiet! Dark Rock Pro 6	63 Max °C, 61 Avg °C
Thermalright Peerless Assassin 120 SE	64 Max °C, 58 Avg °C

Max °C

Avg °C

Arctic Freezer 36-S A-RGB

80

76

Arctic Freezer 36

79

75

Be Quiet! Dark Rock 6

83

78

Be Quiet! Dark Rock 5

90

85

Be Quiet! Dark Rock Pro 6

82

82

Thermalright Peerless Assassin 120 SE

80

71

0255075100

Prime95 @ 1.325 V | Normalised to 20 °C | ~220 W Data

Product	Value
Arctic Freezer 36-S A-RGB	80 Max °C, 76 Avg °C
Arctic Freezer 36	79 Max °C, 75 Avg °C
Be Quiet! Dark Rock 6	83 Max °C, 78 Avg °C
Be Quiet! Dark Rock 5	90 Max °C, 85 Avg °C
Be Quiet! Dark Rock Pro 6	82 Max °C, 82 Avg °C
Thermalright Peerless Assassin 120 SE	80 Max °C, 71 Avg °C

Arctic Freezer 36-S A-RGB

23

Arctic Freezer 36

27

Be Quiet! Dark Rock 6

20

Be Quiet! Dark Rock 5

21

Be Quiet! Dark Rock Pro 6

20

Thermalright Peerless Assassin 120 SE

22

07.51522.530

Ambient temperature Data

Product	Value
Arctic Freezer 36-S A-RGB	23
Arctic Freezer 36	27
Be Quiet! Dark Rock 6	20
Be Quiet! Dark Rock 5	21
Be Quiet! Dark Rock Pro 6	20
Thermalright Peerless Assassin 120 SE	22

We measure the package temperature of a Core i7 14700K across various applications and games at three fixed voltages. This is normalised to an ambient temperature of 20 ℃ for consistency.

I was expecting a bigger delta between the Arctic Freezer 36-S and the bigger, dual-fan Freezer 36. There's not too much in it, just a couple of degrees at most during the more intensive tests.

I test each air cooler using a Core i7 14700K, running at three different fixed voltages: 1.275 V, 1.325 V, and 1.4 V. The highest of which is admittedly a bit of a torture test, running Prime95, which often leads to thermal throttling with an air cooler.

Our game of choice for testing, Baldur's Gate 3, is a little more CPU intensive than some. Though not enough to see huge swings in temperature between coolers at 1.275 V. Prime95 and X264 are very intensive, however, and the Freezer 36-S is second-worst out of the lot in these tests, behind only the Be Quiet! Dark Rock 5. Though that's a more expensive cooler than the Freezer 36-S, and there's not much in it between the Freezer 36-S and the Freezer 36. In Prime95 at 1.325 V, both oddly reached the same temperature, likely a sign of the P12 Pro and its impressive cooling ability at max RPM, and the expected variance in my test process. These aren't lab conditions, folks—as damned near to as I can get in my home office.

Aesthetics

I'm reviewing the A-RGB version of the Freezer 36-S, which includes a P12 Pro A-RGB fan. It's not the prettiest implementation, it must be said, with clear blades and LEDs embedded in the hub in the age-old fashion. But it does produce a pretty enough effect for the price. The RGB effect can be controlled via any 5V RGB header.

The Freezer 36-S looks pretty great underneath the RGB lighting. The version I have for review comes with an all-black coating, which includes heatsink, heatpipes, and even fan clips. It doesn't scream 'budget' like some others do. I wouldn't assume it's a budget model.

Sound

I test each cooler with their fans set to 100% PWM duty cycle, to see what they're capable of. I wouldn't recommend you do the same at home, especially not with the Freezer 36-S. It's very loud, coming in at 47 dBA, and much louder than the Freezer 36 with two slower-spinning fans.

In defence of the Freezer 36-S, it's still quieter than many other coolers I've tested.

Value

Arctic is its own greatest enemy in terms of value. In the US, the Freezer 36 (non-S) is $9 more, at $29, than the Freezer 36-S at $20 (on sale at time of writing, usually $22). In the UK, the Freezer 36 costs just one pound more (£20) than the Freezer 36 (£19), or £4 more for the all-black model to match. Why opt for the smaller, weaker option with one fewer fan when the better pick is only a quick search for spare change down the back of the sofa away?

The Freezer 36 is a tough act to follow, and while prices may fluctuate in favour of the 36-S at times, I just can't see myself buying one—no matter how mighty it might be for the money. There's not much wiggle room here for the Freezer 36, either, as the excellent and enormous Thermalright Peerless Assassin has the market all wrapped up at $40/£36.

It's clear the new MSI Claw 8 EX AI+ CG3EM is the best gaming handheld in terms of almost all the metrics that count. It's the most powerful—delivering gaming performance well beyond that of the ROG Xbox Ally X with its AMD-powered Z2 Extreme APU—and it has an absolutely killer battery life, too. But these aren't the only metrics that count. Sure, they're the key ones if we're talking about what it's like to actually live with the MSI Claw 8 etc. but there's a large hurdle to get over before then. And it's a hurdle built out of pure cash.

The Claw 8 etc. is a handheld that's become as much of a casualty of the current memory and hardware crisis as the Steam Machine. To the point where, here at launch it's just looking like a near $2,000 toy for the really rather rich.

And this is a real failing for a handheld gaming PC. These aren't devices that are meant to be your sole machine; they're complementary systems for when you're travelling around, or just want to sit on a sofa instead of at your gaming desk. Which is why the Steam Deck was such a success. It understood the balance of price and performance was vital in this burgeoning market, and that it needed to be something additive to your gaming experience, not exclusory. And other players, such as Asus and Lenovo, were able to use their buying power to keep costs relatively low for their own subsequent devices.

That's all out of the window now, because there is no way to produce an affordable version of a handheld rocking the outstanding Intel Arc G3 Extreme silicon, whatever your buying power is. Which is a huge shame, because it really is the best gaming handheld I've ever spent time with. The chip at its heart is almost the entire reason behind that; because of its mix of extreme power and extreme Panther Lake efficiency it nails what I've always wanted a handheld gaming PC to be.

The level of compromise over the visual quality and frame rate of your games is impressively light, and while range anxiety is not necessarily a solved issue, the Endurance mode battery life is a bit of a revelation. But it's not all about Intel's silicon, because the enhanced MSI Claw design also plays its part. The weight is beautifully distributed, making it feel far lighter than it should, and the extended controller grips make it far more comfortable for a long session.

MSI Claw 8 EX AI+ features

The key feature for the MSI Claw 8 EX AI+ CG3EM is its Intel Arc G3 Extreme silicon. This is the first chip from Intel that is explicitly designed for a handheld ecosystem, and the fact that it is denoted as an 'Arc' chip highlights its gaming credentials, and the focus on its integrated GPU. In fact, Intel's Tom Petersen went so far as to pitch his claim to actually call the chip a GPU with an integrated CPU when I spoke to him at Computex earlier in the year.

Though I'm pretty confident his pitch didn't get very far with the Intel leadership.

The integrated graphics are the same as those sitting in the top-end Panther Lake processors, the Arc B390. That's a 12 Xe core chip and one that's impressively capable at 1080p gaming, especially at low wattage levels, but the chip as a whole contains some interesting optimisations dedicated to handheld gaming specifically.

The most obvious is how the split between Performance, Efficient, and Low Power Efficient cores lines up in the new chip spec. Where your standard Core UItra 9 or 7 Panther Lake chip comes with four Cougar Cove Performance cores to give it that extra processing grunt, Intel has cut that in half, with only two Performance cores in the G3 series of chips.

Alongside that pair, are eight Darkmont Efficient cores and a further four Darkmont LP Efficient cores. That then gives you a 14-core processor, so you can have some decent multi-core performance in there without dedicating too much silicon to the highest-performing—and most thirsty—cores Intel has in the whole Panther Lake stack.

And the Darkmont cores are plenty versatile and powerful enough in this form, and not the weakheart cores you might have in your head from the first flush of Intel's big/little core design. This isn't a mobile workstation after all, this is a gaming chip, and everything is focused around that.

To that end the G3 chips come with the latest version of Intel's Intelligent Bias Control software (IBC v3.5). This is a piece of software designed to balance power between CPU and GPU tiles, and here it allows for the complete parking of the more power-hungry Performance cores when the chip is operating at a low power level, below 14 W. It also affects the Windows scheduler and points more work towards the Efficient cores; it's almost the entirely opposite way the original scheduler was designed when Intel first brought the hybrid core layout to the PC.

But, having played with a couple of other Arc G3 Extreme handhelds—the OneXPlayer 3 and the Acer Predator Atlas 8—there is a lot that MSI is bringing to the table here, too. And much of it is a vast improvement over the previous Claw handhelds. MSI has been making Intel-based handheld gaming PCs for a couple generations now and they've been decent, but never class-leading. That's changed here with the Claw 8 EX AI+, where it's focused on nailing the design in terms of ergonomics and the ephemeral feel of a device, as well as supporting the efficient Intel architecture with a fat 77.8 Whr battery.

The flared design of the grips mirrors the Xbox controller and therefore the comfort of the Xbox Ally X, and all the buttons feel far more definite than the slightly woolly, imprecise controls on the previous Claw devices. We also have Hall effect sticks and triggers to give you more accuracy and remove the issue of stick drift over time, too.

The 8-inch IPS-like panel taking up the front of the Claw—and oddly slightly jutting out at the bottom—is a really solid 1920 x 1200 screen. With VRR support, you're getting a refresh rate from 48 - 120 Hz and an impressively bright display, too. It's rated for a typical luminance level of 500 nits, which means you don't have to have it anywhere near 100% to get a very playable level of brightness on the thing.

It's not all on the outside, either. The internals have been given an upgrade, and arguably the most important is the one you'll barely notice. The cooling. MSI has used taller fans in the twin cooling array inside it, and there is an impressive amount of clearance between the outer shell and those li'l spinny things. All told, you're getting an extra 5 W of cooling capacity compared to previous Claw designs, and apparently 9% more airflow.

Combine that with the noted Panther Lake/Arc G3 efficiency, and you get a cooling system that stays mercifully quiet, but still impressively effective when you're really pushing the graphics silicon to its fullest.

Another internal feature that pleases me is the SSD. It's super easy to unscrew and pry the back off the Claw—without breaking the plastic shell—and you're immediately presented with the SSD. This makes it simple to replace or upgrade the standard 1 TB of storage it comes with, but MSI has also ensured compatibility with full-size 2280 SSDs as well as the smaller designs.

MSI Claw 8 EX AI+ performance

If you've seen anything of our coverage of Intel's Panther Lake you'll know we're big fans of its gaming performance, and that hasn't changed with my testing of this new MSI Claw and its Arc G3 Extreme processor. In fact, it has only been enhanced because of its effectiveness at low power especially.

Our standard handheld gaming PC benchmarking suite sees the MSI Claw 8 EX AI+ topping the table with only one exception, the ageing Horizon Zero Dawn upscaling benchmark, which relies on an older version of AMD's FSR technology. It's an impressive display, with it not being at all close in most games compared with the current pinnacle of AMD's own dedicated handheld silicon, the Ryzen Z2 Extreme in the ROG Xbox Ally X.

Previous

1 / 3

Next

1080p gaming performance

Black Myth Wukong (1080p Medium)Cyberpunk 2077 (1080p Medium)F1 24 (1080p Medium)Horizon Zero Dawn (1080p Original)Metro Exodus Enhanced (1080p High)

Avg FPS

1% Low FPS

MSI Claw 8 EX AI+ CG3EM

33

16

Asus ROG Xbox Ally X

26

19

MSI Claw 8 AI+

20

15

Lenovo Legion Go S

15

11

Black Myth Wukong (1080p Medium) Data

Product	Value
MSI Claw 8 EX AI+ CG3EM	33 Avg FPS, 16 1% Low FPS
Asus ROG Xbox Ally X	26 Avg FPS, 19 1% Low FPS
MSI Claw 8 AI+	20 Avg FPS, 15 1% Low FPS
Lenovo Legion Go S	15 Avg FPS, 11 1% Low FPS

Avg FPS

1% Low FPS

MSI Claw 8 EX AI+ CG3EM

51

33

Asus ROG Xbox Ally X

41

34

MSI Claw 8 AI+

37

31

Lenovo Legion Go S

22

19

Cyberpunk 2077 (1080p Medium) Data

Product	Value
MSI Claw 8 EX AI+ CG3EM	51 Avg FPS, 33 1% Low FPS
Asus ROG Xbox Ally X	41 Avg FPS, 34 1% Low FPS
MSI Claw 8 AI+	37 Avg FPS, 31 1% Low FPS
Lenovo Legion Go S	22 Avg FPS, 19 1% Low FPS

Avg FPS

1% Low FPS

MSI Claw 8 EX AI+ CG3EM

107

64

Asus ROG Xbox Ally X

84

76

MSI Claw 8 AI+

68

59

Lenovo Legion Go S

45

34

F1 24 (1080p Medium) Data

Product	Value
MSI Claw 8 EX AI+ CG3EM	107 Avg FPS, 64 1% Low FPS
Asus ROG Xbox Ally X	84 Avg FPS, 76 1% Low FPS
MSI Claw 8 AI+	68 Avg FPS, 59 1% Low FPS
Lenovo Legion Go S	45 Avg FPS, 34 1% Low FPS

Avg FPS

1% Low FPS

MSI Claw 8 EX AI+ CG3EM

61

32

Asus ROG Xbox Ally X

48

15

MSI Claw 8 AI+

48

11

Lenovo Legion Go S

36

3

Horizon Zero Dawn (1080p Original) Data

Product	Value
MSI Claw 8 EX AI+ CG3EM	61 Avg FPS, 32 1% Low FPS
Asus ROG Xbox Ally X	48 Avg FPS, 15 1% Low FPS
MSI Claw 8 AI+	48 Avg FPS, 11 1% Low FPS
Lenovo Legion Go S	36 Avg FPS, 3 1% Low FPS

Avg FPS

1% Low FPS

MSI Claw 8 EX AI+ CG3EM

41

15

Asus ROG Xbox Ally X

27

15

MSI Claw 8 AI+

25

17

Lenovo Legion Go S

14

8

Metro Exodus Enhanced (1080p High) Data

Product	Value
MSI Claw 8 EX AI+ CG3EM	41 Avg FPS, 15 1% Low FPS
Asus ROG Xbox Ally X	27 Avg FPS, 15 1% Low FPS
MSI Claw 8 AI+	25 Avg FPS, 17 1% Low FPS
Lenovo Legion Go S	14 Avg FPS, 8 1% Low FPS

The Arc G3 Extreme chip in the Claw 8 EX AI+ shows how dominant it is over AMD's best handheld chip, and the last generation Intel iGPU in the previous MSI Claw 8 device.

1080p upscaling performance

Black Myth Wukong | Upscaling (1080p Medium)Cyberpunk 2077 | Upscaling (1080p Medium)F1 24 | Upscaling (1080p Medium)Horizon Zero Dawn | Upscaling (1080p Original)

Avg FPS

1% Low FPS

MSI Claw 8 EX AI+ CG3EM

50

28

Asus ROG Xbox Ally X

39

30

MSI Claw 8 AI+

33

25

Lenovo Legion Go S

22

18

Black Myth Wukong | Upscaling (1080p Medium) Data

Product	Value
MSI Claw 8 EX AI+ CG3EM	50 Avg FPS, 28 1% Low FPS
Asus ROG Xbox Ally X	39 Avg FPS, 30 1% Low FPS
MSI Claw 8 AI+	33 Avg FPS, 25 1% Low FPS
Lenovo Legion Go S	22 Avg FPS, 18 1% Low FPS

Avg FPS

1% Low FPS

MSI Claw 8 EX AI+ CG3EM

73

50

Asus ROG Xbox Ally X

53

44

MSI Claw 8 AI+

48

40

Lenovo Legion Go S

29

25

Cyberpunk 2077 | Upscaling (1080p Medium) Data

Product	Value
MSI Claw 8 EX AI+ CG3EM	73 Avg FPS, 50 1% Low FPS
Asus ROG Xbox Ally X	53 Avg FPS, 44 1% Low FPS
MSI Claw 8 AI+	48 Avg FPS, 40 1% Low FPS
Lenovo Legion Go S	29 Avg FPS, 25 1% Low FPS

Avg FPS

1% Low FPS

MSI Claw 8 EX AI+ CG3EM

124

88

Asus ROG Xbox Ally X

97

88

MSI Claw 8 AI+

85

75

Lenovo Legion Go S

57

21

F1 24 | Upscaling (1080p Medium) Data

Product	Value
MSI Claw 8 EX AI+ CG3EM	124 Avg FPS, 88 1% Low FPS
Asus ROG Xbox Ally X	97 Avg FPS, 88 1% Low FPS
MSI Claw 8 AI+	85 Avg FPS, 75 1% Low FPS
Lenovo Legion Go S	57 Avg FPS, 21 1% Low FPS

Avg FPS

1% Low FPS

MSI Claw 8 EX AI+ CG3EM

66

51

Asus ROG Xbox Ally X

70

18

MSI Claw 8 AI+

54

23

Lenovo Legion Go S

51

8

Horizon Zero Dawn | Upscaling (1080p Original) Data

Product	Value
MSI Claw 8 EX AI+ CG3EM	66 Avg FPS, 51 1% Low FPS
Asus ROG Xbox Ally X	70 Avg FPS, 18 1% Low FPS
MSI Claw 8 AI+	54 Avg FPS, 23 1% Low FPS
Lenovo Legion Go S	51 Avg FPS, 8 1% Low FPS

The Arc G3 Extreme chip in the Claw 8 EX AI+ shows how dominant it is over AMD's best handheld chip, and the last generation Intel iGPU in the previous MSI Claw 8 device.

Synthetic performance

3DMark Time Spy3DMark StorageFFXIV Shadowbringers

Index score

GPU index score

CPU index score

MSI Claw 8 EX AI+ CG3EM

6499

6091

10487

Asus ROG Xbox Ally X

4033

3663

9462

MSI Claw 8 AI+

4460

4125

8266

Lenovo Legion Go S

2244

2072

4246

3DMark Time Spy Data

Product	Value
MSI Claw 8 EX AI+ CG3EM	6499 Index score, 6091 GPU index score, 10487 CPU index score
Asus ROG Xbox Ally X	4033 Index score, 3663 GPU index score, 9462 CPU index score
MSI Claw 8 AI+	4460 Index score, 4125 GPU index score, 8266 CPU index score
Lenovo Legion Go S	2244 Index score, 2072 GPU index score, 4246 CPU index score

Index score

Avg bandwidth (MB/s)

Access time (ųs)

MSI Claw 8 EX AI+ CG3EM

2152

366.66

83

Asus ROG Xbox Ally X

1163

199.98

155

MSI Claw 8 AI+

1889

323.51

95

Lenovo Legion Go S

1582

272.24

114

3DMark Storage Data

Product	Value
MSI Claw 8 EX AI+ CG3EM	2152 Index score, 366.66 Avg bandwidth (MB/s), 83 Access time (ųs)
Asus ROG Xbox Ally X	1163 Index score, 199.98 Avg bandwidth (MB/s), 155 Access time (ųs)
MSI Claw 8 AI+	1889 Index score, 323.51 Avg bandwidth (MB/s), 95 Access time (ųs)
Lenovo Legion Go S	1582 Index score, 272.24 Avg bandwidth (MB/s), 114 Access time (ųs)

MSI Claw 8 EX AI+ CG3EM

10.952

Asus ROG Xbox Ally X

18.91

MSI Claw 8 AI+

15.15

Lenovo Legion Go S

20.56

FFXIV Shadowbringers Data

Product	Value
MSI Claw 8 EX AI+ CG3EM	10.952
Asus ROG Xbox Ally X	18.91
MSI Claw 8 AI+	15.15
Lenovo Legion Go S	20.56

The Arc G3 Extreme chip in the Claw 8 EX AI+ shows how dominant it is over AMD's best handheld chip, and the last generation Intel iGPU in the previous MSI Claw 8 device.

As soon as you start twiddling the upscaling knobs you're getting to never less than fully playable on all the games in our tests. And that's without even going down the frame generation route. If you find a game, such as Lego Batman: Legacy of the Dark Knight, which has Intel's own XeSS FG baked into it, you are laughing. It pairs so well with the Arc G3 Extreme chip that Intel believes it should become a vital part of your gaming experience, and I have to agree.

From spending time playing Batman on the device, with multi frame gen enabled, it really does make a huge difference to the smoothness of the game, without hugely impacting the input latency. It's one of those times where I completely forgot what I was testing because all of the tech and frame gen faded into the background as I just spent time playing a great PC game without thinking about anything else.

Thankfully, that's how Intel is talking about frame gen with regards to the G3 Extreme series. It's all about making the experience smoother, not necessarily about winning people over with explicitly higher measured frame rates. That is what frame generation is good for: smoothing out gaming performance, not for giving you a veneer of higher performance.

Previous

1 / 2

Next

Endurance mode performance

Black Myth Wukong | Upscaling (1080p Medium)Cyberpunk 2077 | Upscaling (1080p Medium)F1 24 | Upscaling (1080p Medium)Horizon Zero Dawn | Upscaling (1080p Original)Metro Exodus Enhanced (1080p High)

Avg FPS

1% Low FPS

MSI Claw 8 EX AI+ CG3EM (Endurance mode)

37

29

MSI Claw 8 EX AI+ CG3EM (Default)

50

28

Asus ROG Xbox Ally X (Default)

39

30

Black Myth Wukong | Upscaling (1080p Medium) Data

Product	Value
MSI Claw 8 EX AI+ CG3EM (Endurance mode)	37 Avg FPS, 29 1% Low FPS
MSI Claw 8 EX AI+ CG3EM (Default)	50 Avg FPS, 28 1% Low FPS
Asus ROG Xbox Ally X (Default)	39 Avg FPS, 30 1% Low FPS

Avg FPS

1% Low FPS

MSI Claw 8 EX AI+ CG3EM (Endurance mode)

51

32

MSI Claw 8 EX AI+ CG3EM (Default)

73

50

Asus ROG Xbox Ally X (Default)

53

44

Cyberpunk 2077 | Upscaling (1080p Medium) Data

Product	Value
MSI Claw 8 EX AI+ CG3EM (Endurance mode)	51 Avg FPS, 32 1% Low FPS
MSI Claw 8 EX AI+ CG3EM (Default)	73 Avg FPS, 50 1% Low FPS
Asus ROG Xbox Ally X (Default)	53 Avg FPS, 44 1% Low FPS

Avg FPS

1% Low FPS

MSI Claw 8 EX AI+ CG3EM (Endurance mode)

87

59

MSI Claw 8 EX AI+ CG3EM (Default)

124

88

Asus ROG Xbox Ally X (Default)

97

88

F1 24 | Upscaling (1080p Medium) Data

Product	Value
MSI Claw 8 EX AI+ CG3EM (Endurance mode)	87 Avg FPS, 59 1% Low FPS
MSI Claw 8 EX AI+ CG3EM (Default)	124 Avg FPS, 88 1% Low FPS
Asus ROG Xbox Ally X (Default)	97 Avg FPS, 88 1% Low FPS

Avg FPS

1% Low FPS

MSI Claw 8 EX AI+ CG3EM (Endurance mode)

47

36

MSI Claw 8 EX AI+ CG3EM (Default)

66

51

Asus ROG Xbox Ally X (Default)

70

18

Horizon Zero Dawn | Upscaling (1080p Original) Data

Product	Value
MSI Claw 8 EX AI+ CG3EM (Endurance mode)	47 Avg FPS, 36 1% Low FPS
MSI Claw 8 EX AI+ CG3EM (Default)	66 Avg FPS, 51 1% Low FPS
Asus ROG Xbox Ally X (Default)	70 Avg FPS, 18 1% Low FPS

Avg FPS

1% Low FPS

MSI Claw 8 EX AI+ CG3EM (Endurance mode)

20

15

MSI Claw 8 EX AI+ CG3EM (Default)

41

15

Asus ROG Xbox Ally X (Default)

27

15

Metro Exodus Enhanced (1080p High) Data

Product	Value
MSI Claw 8 EX AI+ CG3EM (Endurance mode)	20 Avg FPS, 15 1% Low FPS
MSI Claw 8 EX AI+ CG3EM (Default)	41 Avg FPS, 15 1% Low FPS
Asus ROG Xbox Ally X (Default)	27 Avg FPS, 15 1% Low FPS

With the MSI Claw running in its Endurance mode you are getting impressively close to the full performance of the AMD Z2 Extreme chip in Xbox Ally X.

Battery life

PCMark 10 | Battery life (gaming test)

MSI Claw 8 EX AI+ CG3EM (Endurance mode)

308

MSI Claw 8 EX AI+ CG3EM (Default)

153

Asus ROG Xbox Ally X (Default)

125

MSI Claw 8 AI+ (Default)

129

Lenovo Legion Go S (Default)

126

0100200300400

Battery life (minutes)

PCMark 10 | Battery life (gaming test) Data

Product	Value
MSI Claw 8 EX AI+ CG3EM (Endurance mode)	308
MSI Claw 8 EX AI+ CG3EM (Default)	153
Asus ROG Xbox Ally X (Default)	125
MSI Claw 8 AI+ (Default)	129
Lenovo Legion Go S (Default)	126

With Endurance mode you are literally doubling the battery life of the MSI Claw compared with its already impressive standard showing.

But one of the things that really stands out to me is the efficiency of the new Intel silicon when it comes to Endurance mode. When you enable Endurance in the MSI quick settings it turns on efficiency mode, which theoretically locks the frame rate to 30, 40, or 60 fps. Well, I'm not entirely sure about that, given the variety in average fps I've measured in my benchmarking, but it certainly pulls down the peak PL2 performance down to a maximum of 20 W, with PL1 stuck at 15 W.

This effectively doubles the battery life of the device, but delivers gaming performance that, to all practical purposes, feels about the same as you would get from running the Xbox Ally X at its top settings. There are times where the full power of the Z2 Extreme does edge out the 15 - 20 W performance of the MSI Claw, but those aren't all that common.

This performance and efficiency one-two punch is a hell of a double-whammy.

MSI Claw 8 EX AI+ value

Value? In the current handheld gaming market, surely you jest? This is where all my positivity about the new Intel Arc G3 series architecture and MSI's re-design of the Claw evaporates up against the molten heat of the memory crisis, because there is no getting away from the fact that we're talking about a gaming handheld that costs $1,800.

As soon as the AI/datacentre market started hoovering up all available DRAM and subsequently NAND memory the writing was on the wall for consumer technology. The laws of supply and demand meant prices were only going to spiral upwards and this has arguably hit the MSI Claw harder than many others.

That's because Panther Lake and the Arc G3 series chips' performance targets were designed around the bandwidth of high-performance LPDDR5X memory. Intel reportedly mandated at least LPDDR5X-7467 for a system to be classed as having either the top B370 or B390 iGPUs, otherwise they would just be noted as 'Intel Graphics'. The G3 Extreme here sits alongside 32 GB of LPDDR5X-8533, which costs a pretty penny and is likely a big reason for the high price of the device itself.

I spoke to Intel's Tom Petersen at Computex earlier this year specifically about whether we'd see any machines using the non-Extreme version of the two G3 chips the company announced, and he was sceptical precisely because of this.

"I don't know the answer to that one," he says. "It's a great question. I think we're in a weird time with the whole DRAM thing going on right now, where the pricing difference between the two chips is small, and small relative to what's going on with DRAM.

"So, if you're going to spend the money on DRAM, why not just get the best one? I think there's some of that going on. As DRAM prices come down and the market starts normalising a little bit, then the differentiation of price that you can get by picking different SKUs will be a more important differentiator than the end user. That's my guess."

Which kinda sucks. The memory just costs far more than the difference between the two G3 processors, making that saving largely irrelevant for anyone looking to try and create a cheaper option.

You only have to look at the Acer Predator Atlas 8, which is also using the G3 series of chips and is releasing a straight Arc G3 version. But that's looking to be priced around $1,500 despite using the cheaper chip. The equivalent Atlas 8 compared to the MSI Claw 8 etc. is running to over $2,000. However you cut it, the mix of CPU and memory silicon is expensive.

So, the price is understandable then, but that doesn't change the fact that at this level it's impossible to recommend the Claw 8 etc. to anyone who makes their purchases with any form of value judgement in mind. If you're just after the best/fastest, and hang the expense, all power to you, but for the rest of us there's little by way of a value proposition here.

MSI Claw 8 EX AI+ sound

When it comes to the actual sound reproduction of the speakers in the MSI Claw EX AI+ they're fine. Nothing more, nothing less. Arguably there's not a whole lot you can do with the sort of space you have to jam a set of speakers into, but they still do an admirable job anyway. Though I will say that I'd still advise you to use a set of earbuds or headphones with the device for the best aural experience.

But the other side of this equation is the sound the device's cooling makes when you're pushing the silicon to its limits. And in that I'm far more impressed.

It's not that the device is silent under load—you will still hear those fans when they're really blowing—but the tone is as unobtrusive as they get and the work MSI has done to improve the cooling array of the latest MSI Claw has really paid off. If you're rocking Endurance mode and playing some relatively lightweight games you will barely hear the whoosh of air carrying off the heat from the silicon.

And even if you are hammering the G3 Extreme, I never felt the same sort of loathing I get from a really egregious gaming laptop's turbine roar under gaming loads.

MSI Claw 8 EX AI+ thermals

This is an Intel CPU, so when we really put it under sustained load, such as with a ten minute Cinebench 2024 run, then yes, we're going to be hitting high temperatures. But it's also no different to the sorts of temperatures you'll see from AMD's own handheld chips when put under the same processing pressure.

There's a solitary degree in it, which could easily be passed off as variance in testing, but the Arc G3 Extreme in the MSI Claw measured 97 °C as its maximum temperature, while the Ryzen Z2 Extreme in the Asus ROG Xbox Ally X hits 96 °C.

Cooling

Thermal performance

Max CPU temp (°C)

Max CPU package power (W)

MSI Claw 8 EX AI+ CG3EM

97

37

Asus ROG Xbox Ally X

96

55

Lenovo Legion Go S

99.2

38

Thermal performance Data

Product	Value
MSI Claw 8 EX AI+ CG3EM	97 Max CPU temp (°C), 37 Max CPU package power (W)
Asus ROG Xbox Ally X	96 Max CPU temp (°C), 55 Max CPU package power (W)
Lenovo Legion Go S	99.2 Max CPU temp (°C), 38 Max CPU package power (W)

The thermal performance of the MSI Claw is impressive, however the Xbox Ally X is doing more work with higher power draw.

But at no point do either of these handhelds ever get to the point where they are too hot to touch. The cooling design vents the hot air out of the top of the MSI Claw, while drawing it in from twin fans oriented to pull in from the rear of the device. This means you're never hit by the hot air when you're using it in a normal way, with the screen out in front of you.

MSI Claw 8 EX AI+ software

We're still talking about a Windows-based handheld, and even with the Xbox Mode operating to full effect, the MSI Claw remains a device that feels like it's got a bunch of software snakes writhing around under the surface all fighting to control what happens in there. The simple, unified nature of SteamOS is what has drawn people to Valve's Linux-based operating system, and why a bunch of other Linux distributions have co-opted it as the front-end for their own handheld versions.

But Windows is still very much a desktop OS, and one that still has some roots in the aging operating systems of its past that it cannot escape. There are still Windows 2000-era menu systems lurking inside Windows 11 if you look not very hard at all. And even with the Xbox mode layer cleaning things up for handhelds it's still not entirely unified.

Yes, you can have Steam boot into Big Picture mode, but I found Xbox mode seemed to block that from loading immediately after booting. So you're still going to boot into either MSI's own game and system launcher or the Xbox app. Either way is clunky and requires you to jump through a few more hoops, click a few more buttons, to get to the game you want.

Then there's the overlay. Pressing the menu button on the Claw brings up what's essentially the Gamebar menu as an overlay, just like with the Xbox Ally X. Except here, after maybe a one second delay, it will jarringly flick you over to the MSI quick menu. I will say this is probably where you want to land when you hit that button, as it gives you all the necessary control over performance, controls, and recording, but it's not a slick experience.

MSI Claw 8 EX AI+ battery life

If gaming performance is the attention grabbing metric for most PC gamers when it comes to handheld PCs, battery life is the one that will have the biggest impact on your overall experience with the device when you get it in your hands.

I was already impressed with the MSI Claw EX AI+ CG3EM when I first ran my battery life benchmarks on its default, AI Engine power mode. This gets you running up to 37 W and I am getting another half an hour of solid gaming time over and above the Xbox Ally X, and that's all while performing at a much higher level.

But the joy for me of the new Arc G3 Extreme chip is its performance when you're not throwing that much power down its throat. Switch on Endurance mode in the MSI quick menu, which puts you in the Efficiency preset in the Intel Graphics Software app, and my measured battery life literally doubles.

Battery life

PCMark 10 | Battery life (gaming test)

MSI Claw 8 EX AI+ CG3EM (Endurance mode)

308

MSI Claw 8 EX AI+ CG3EM (Default)

153

Asus ROG Xbox Ally X (Default)

125

MSI Claw 8 AI+ (Default)

129

Lenovo Legion Go S (Default)

126

Asus ROG Ally X (Default)

128

OneXPlayer OneXFly F1 Pro (Default)

82

0100200300400

Battery life (minutes)

PCMark 10 | Battery life (gaming test) Data

Product	Value
MSI Claw 8 EX AI+ CG3EM (Endurance mode)	308
MSI Claw 8 EX AI+ CG3EM (Default)	153
Asus ROG Xbox Ally X (Default)	125
MSI Claw 8 AI+ (Default)	129
Lenovo Legion Go S (Default)	126
Asus ROG Ally X (Default)	128
OneXPlayer OneXFly F1 Pro (Default)	82

We use PCMark's Gaming battery life test, which loops through older 3DMark tests, continually hammering the battery until it gives out and the system goes into shutdown. And with Endurance mode enabled I've seen the Claw go from two and a half hours in that test to over five hours. And, remember, that will give you essentially the same actual gaming performance as the ROG Xbox Ally X across most games I've tested.

Now, that doesn't mean battery life is a solved issue, but this is you playing games without worrying about what that little percentage indicator in the top right of your screen says. And if you're playing something like Olli Olli World or Lonely Mountains: Snow Riders, you can expect to bat well above that five hours average, too.

In fact, Intel actually measured the battery life of the Claw with the classic Team Fortress 2 running on it, and managed to get a healthy 11 hours of game time out of it. Actually, nearer 12. That's more than enough TF2 for anyone.

Glorious, purveyor of PC peripherals such as gaming keyboards and mice, has branched out into headsets. The company released its first wired effort, the Glorious GHS Eternal, earlier this year and, in terms of build quality, the wireless InfinitePlay shares a lot of DNA. In terms of pricing though, the two headsets couldn't be more different.

The swappable battery system goes some way to justify the significant jump in price. As someone who loves wireless headsets, but doesn't love how often they have to become tethered to my PC to charge, this proved a real game-changer.

The InfinitePlay is also designed to be a versatile bit of kit you can take everywhere. Whether you're listening to tunes on the go, or counting the shots in one of the best FPS games, the InfinitePlay's balanced sound profile is tuned to handle both scenarios equally well. EQ tweaking is offered via the straightforward to use Glorious Core software, with more in-depth options planned for a future update.

That versatility is backed up by great connectivity. You've got the usual choice between a Bluetooth, or 2.4 GHz connection, and the dongle for the latter comes with a USB-C to USB-A adapter. As such, it's just as easy to connect the InfinitePlay to your phone, PlayStation and Nintendo Switch consoles as it is to your PC.

Generally speaking, the GHS InfinitePlay is a wireless gaming headset that could handle everything I threw at it—and, thanks to its sturdy yet flexible construction, survived everywhere I threw it. Unfortunately, for $230, I'm not sold that this headset gives quite as good as it gets.

Features

The main draw of the Infiniteplay is its swappable battery system. Each battery holds an advertised 85 hours of charge while the headset itself can stay powered on for three hours under its own steam. That means, your headset doesn't need to power off when you swap between batteries

We've seen similar systems in both the SteelSeries Arctis Nova Elite, and the SteelSeries Arctis Nova Pro. The InfinitePlay costs a fraction of the Nova Elite's $600 asking price, and a significant chunk of change less than the Nova Pro, which itself will still set you back somewhere in the region of $300. The InfinitePlay's $230 price tag still isn't what I'd call affordable, but it does offer more than twice the battery life of the more expensive SteelSeries headsets—seriously, I was charging the Glorious batteries in their special included base station maybe once every two weeks.

Swapping between batteries feels natural. There's no hidden compartment here, and therefore no magnetic cap that could disappear among the clutter of your gamer den. Instead, you just slot a freshly charged battery into the right earcup. It does matter which end of the battery goes in first, with the non-business end highlighted by an orange tip. Otherwise, the socket is much less picky about which side is the right way up, so you won't be futzing and flipping your batteries around like they're a USB-A cable. Easy-peasy.

Besides hot swappable batteries, or the impressive, detachable microphone which I'll discuss in its own section, what I personally most enjoyed were the tactile rotary knobs on both earcups. Pressing in the left knob powers on the headset, while tapping gives you a broad sense of how much juice is left in the battery. Twiddling the left knob also finetunes your chat mix, while twisting rightie adjusts system volume. Tapping the right-hand dial also pauses the current song.

The two rotary dials are a delight to use, and each one is joined by a lower profile button lower down the earcup. The left ear cup button mutes the mic, and the one on the right earcup turns on Bluetooth with a long press, or switches between your Bluetooth and 2.4 GHz connection with a tap. These lower profile buttons are fine, though they're much less fun to use.

Sound

The InfinitePlay enjoys a balanced sound profile with various EQ tweaking options available via the Glorious Core software (scroll down for more on that). But even without audio twiddling, the InfinitePlay handles both music listening and game soundscapes equally well. However, that balanced sound profile also means it doesn't especially stand out in any particular regard.

That's fine, but for this price point I would want just a little more depth and richness—a little more bass wouldn't hurt either. Instead, Glorious is pitching this as more of a versatile device you can comfortably use both in and out of game, rather than an overly specialised bit of gamer gear.

Still, in-game performance matters. Granted, I could wax lyrical about my daughter's street fighting career in Magical Princess, but the mostly chill soundscape of this raising game doesn't really put the headset through its paces. So, for my yearly rhythm game humbling, I took DJMAX Respect V for a spin, and was particularly impressed with how crisp the high tones sounded via the InfinitePlay. The many voice lines of short, scifi fable Eliza also came through crisp and clear. The funky, crunchy soundscape of Doom: The Dark Ages is also faithfully rendered without the details of a hectic firefight becoming too muddy.

That said, the Glorious plan was to build a versatile bit of hardware with sound you could tune to your liking. As such, I'll save my thoughts on EQ profiles for the software section down below.

Microphone

The InfinitePlay's detachable microphone proves to be another stand-out feature. For instance, when I first tested this wireless headset in a video meeting, one of my colleagues commented that the mic was so clear that it made me sound like I was part of her inner monologue. Beyond considering a career as someone's literal shoulder devil, that is to say this headset could be a real contender for streamers (or, simply anyone just as enamoured with the sound of their own voice).

I'll talk more about the software specifically in a minute, but Glorious Core offers a few settings that bolster the microphone's performance, such as an Environmental Noise Cancellation toggle. Contrary to popular belief, PCG towers isn't the rowdiest office, though this noise filter definitely helps during team video calls.

Because Glorious are pitching the InfinitePlay as a versatile headset meant to handle your auditory needs both in and out of game, you have the choice between letting the detachable boom arm render your dulcet tones, or trusting an internal microphone to keep up. The detachable microphone arm is sturdy, with a little bit of flex as opposed to the properly bendy arms on Glorious' wired headsets.

As I said before, I got the best results using this detachable arm. Detach the boom arm though, and you have a headset that looks much less gamer-y, but with a much less strong internal microphone. It's mostly clear enough, but I wouldn't want to take a phone call on a crowded street or on a windy day with it.

Build quality

Like I said up top, the InfinitePlay is very similar in construction to the Glorious GHS Eternal wired headsets. That means they both feature a fabulously bendy-but-still-sturdy headband. Specifically, the InfinitePlay goes for the same floating headband design as the pricier, RGB version of the Eternal. Much like that wired headset, the InfinitePlay managed to endure a number of trips stuffed into my over-full backpack, coming out no worse for wear on the other side.

Besides being a wireless headset, the InfinitePlay's design differs from the GHS Eternal in two notable ways. For one, it's got two lovely tactile knobs, one on each ear cup, for adjusting volume and chat levels. For another, it does away with its siblings' bendy mic arm and opts for a much less flexible, detachable arm instead. There's some give to this arm, but it's not as poseable, and feels way more robust.

Otherwise, the InfinitePlay's hardware profile is very nearly the same as the GHS Eternal RGB. That means you're getting decent build quality for the price as far as that wired headset is concerned, but standing this side of the $200 mark with the InfinitePlay, the amount of flex possible here may prove more disconcerting than reassuring.

Aesthetics

Glorious haven't yet announced any fun colourways for the InfinitePlay—as a smaller company only just dipping its toe into gaming headsets, that makes sense. What that left me with was a classy black colourway with orange accents on the floating headband and inside the earcups. A white version is also available, though that goes for a grey headband and earcups instead of orange.

In all, my review unit is a fairly understated bit of kit in black. Even the rotary knobs are perfectly in proportion and don't draw a great deal of attention to themself. The microphone arm is a sleek thing, and the option to fully detach it makes this headset an unassuming bit of kit to take on the road with you. Personally, I prefer a versatile bit of kit that also comes in a fetching purple, but I won't begrudge those who prefer a more understated colourway—after all, black does go with everything!

Software

The InfinitePlay comes out of the box with decent microphone audio and a balanced sound profile, but the Glorious Core software offers a number of settings to better hone your experience.

For listening to your favourite tunes (or in my case most recently, going absolutely feral to The Vampire Lestat's singles), the Music and Immersion EQ profiles offer a lovely wide sound stage. You can't tweak the parametrics of either these, or the other Voice Call, FPS, or 'Glorious Mode' pre-set EQ profiles, but Core offers you two custom profiles to futz around with via a fairly intuitive UI.

Doom: The Dark Ages arguably features the weakest soundtrack of the trilogy, but its layered soundscape proves a worthwhile test of any gaming headset. I played the game across a few of Core's pre-loaded EQ profiles, and feel that on the whole both hardware and software are up to the task of faithfully rendering the sound design's redoubled efforts to really sell the heft of every step you make as the Doom Slayer. As far as I can hear, the FPS EQ makes the sound stage slightly narrower compared to the Immersion profile.

Generally though, I didn't notice the biggest difference between the pre-loaded EQ profiles, but I'm told that glorious is planning to beef up its software offering with the addition of advanced EQ options in future Core updates. Per the press release, the plan is to introduce "exact frequency placement, gain and volume per band, Q factor (the width of each adjustment, from narrow and focused to broad and sweeping), and full filter type selection across low pass, high pass, peak, low shelf, and high shelf." Unfortunately, these specific software features were not available to test for my review.

At time of writing, the software offers a number of other options beyond EQ settings too. You can also adjust game and chat sound balance via slider. You can adjust microphone volume, turn sidetone on or off, adjust sidetone volume, and turn the Environmental Noise Cancellation for the microphone on or off too. You can fiddle with all of these and the EQ settings across up to five different headset settings.

Value

That $230 price point makes the InfinitePlay markedly cheaper than a number of competing gaming headsets also rocking swappable battery systems—but dropping 200 smackers on a gaming headset is still a big ask. To reiterate, the Turtle Beach Stealth Pro can still cost more than $300—though our Katie wasn't especially fond of it anyway, giving it a 60% back in 2023. Turtle Beach's successor headset, the Stealth Pro II, offers an improved feature set though is even more expensive, and enjoys only 40 hours of juice per battery. Then, there's the $400 SteelSeries Arctis Nova Pro Omni which is definitely very snazzy, but still only offers up to 60 hours per swappable battery.

If we were just comparing battery life, then the 85 hours offered by each of the InfinitePlay's swappable batteries would ensure the Glorious wins out easily. The headset's own three hours of onboard juice just sweetens the deal, meaning you don't have to power off while swapping batteries. But obviously, wireless gaming headsets are more than just the sum of their battery life.

The audio quality of the detachable microphone is another highlight, but it's not quite enough to justify the price cruising past the $200 mark. Furthermore, while I don't dislike the well-balanced sound from the InfinitePlay 40 mm drivers, I would've appreciated more of an attempt at bass for the price.

Special commentary on this classic PC Gamer re-review provided by:

I’m not sure anybody has ever had more fun making a game than Sumo Digital did when it made Sonic & All-Stars Racing Transformed.

On paper, the brief from SEGA would have been as desperately mercenary as they come: make a copy of Nintendo’s monumentally popular Mario Kart, but with Sonic. Take Nintendo’s near-perfect party game series and smear Sega’s blue weirdo and his friends all over it.

Sumo seems to have understood that it was never going to make something pristine with the materials it was given. SEGA found itself resolutely Mean Girls-ed by the end of the ’90s console wars, too scrappy to mean anything as a brand and personally victimised by Nintendo’s radiant coherence.

As mascots go, Mario is a stranger proposition than Sonic—a portly plumber is a weird hero to invest in when you’re a child—but sheer craftsmanship overruled the rule of cool. Sonic found himself too inherently ’90s to hold onto that cool very long anyway, destined for a faltering retreat from the mainstream and a new career in fan fiction.

The first wonderful thing about All-Stars Racing Transformed is that it accepts this: not just about Sonic, but about SEGA’s whole brand. The core cast includes everybody you’d expect: Sonic, Tails, Knuckles and so on, each driving cars that transform into planes and boats as stages change over the course of a race.

Then somebody put their hands on the crossover lever and gave it a push: you can also play as the bloke from Shenmue, who rides around in an OutRun cabinet, or as B.D. Joe from Crazy Taxi. You can play as AiAi from Super Monkey Ball, who appears to have the Dreamcast logo shaved into his head: a gang sign that only middle class ’90 kids will understand.

Nintendo has done this sort of thing before, of course, in the Smash Bros. series. Hell, Sonic is in Smash Bros. Not content to be outdone, however, Sumo leaned even harder on the crossover lever and snapped it off, wholesale embracing Sega’s latter-day existence as a publisher of sober PC strategy games.

You’d be forgiven for not knowing that you wanted to do sick drifts across the deck of After Burner’s aircraft carrier as a tank-driving general from Company of Heroes: but you do, you really do.

You want to fuck up Tails with a well-aimed swarm of bees as a Roman centurion from Total War. You want to fly around a level from Sonic 3 as actual human racing driver Danica Patrick.

Harvest time

You want to do it all as Racing Transformed’s inarguable star, Football Manager: just a regular man in a football-themed car who likes to bellow things like “get in!” as he tumbles from a collapsing bridge in Panzer Dragoon Saga and transforms into a boat.

I can’t express how happy I am that somebody decided to make this, that they were willing to shatter the ideal of a ’90s console brand into pieces and throw it to the crowd.

Asus isn't the immediately obvious pick when it comes to value-orientated PC hardware. But its mainstream Strix sub-brand does occasionally deliver some pretty sweet deals. Enter, therefore, the new Asus ROG Strix OLED XG32UQWMS.

It's a 32-inch, 4K gaming monitor based on LG's latest Tandem Primary RGB OLED panel tech, the specifics of which we'll come to shortly. And it's yours for $899, which is a pretty punchy price for a 32-inch 4K OLED with the very newest panel tech.

However, as a Strix model, the XG32UQWMS doesn't throw everything at the wall when it comes to features. Well, that's the theory. In practice, the XG32UQWMS has an awful lot of the stuff you'd want from a gaming display.

The one obvious omission is USB-C connectivity. But then that's not an unambiguously gaming-adjacent feature, albeit it is desirable for use with a gaming laptop and likewise if you're planning on sharing this monitor between a gaming rig and, perhaps, a work laptop.

Anyway, apart from the LG-supplied OLED panel itself, there's one little extra that's worth mentioning first. Along with the 240 Hz available at the OLED panel's native 4K resolution, there's also a 1080p dual mode running at fully 480 Hz.

Yes, these kinds of dual mode features are increasingly common on 4K gaming monitors. But given the Strix branding and relatively competitive pricing, not to mention the fact that not many OLED 4K monitors have yet offered it, you still might not have expected it on this particular model.

As for other specifications, this Asus represents a step up in terms of brightness compared with previous 32-inch 4K monitors with LG OLED models. The full-screen rating is now 330 nits, up from 275 nits on earlier models.

Asus is also quoting fully 1,500 nits for peak HDR brightness in a small window. Notably, both of those figures are well up on the 300 nits and 1,000 nits, respectively, that MSI claims for the MPG 322UR X24.

That MSI model uses the latest QD-OLED 32-inch 4K panel from Samsung. So, on paper at least, this new LG panel tech seems to have the edge over Samsung. For what it's worth, both this new Asus and the MSI get VESA DisplayHDR True Black 500 certification. Despite all that, the MSI is quite a bit more expensive, at $1,099.

There are a few other upsides beyond mere brightness that come with this latest LG Tandem OLED generation. First, there's a glossy panel coating. That's not to everyone's taste, but it certainly heightens one of the main strengths of OLED tech, namely contrast.

There's also more colour performance, now up to 99.5% of the DCI-P3 gamut and much more competitive in that regard with Samsung QD-OLED. Asus also says this generation of LG OLED panel tech is 60% more durable in terms of light-emitting lifespan and so should resist burn-in even more robustly, an issue which is covered in any case by the three-year warranty.

What you don't get, however, is true RGB-stripe sub-pixel structure. Confusingly, the "Primary RGB" nomenclature LG applies to this new panel tech refers to the layers of emissive material actually within each subpixel.

The sub-pixel structure itself remains the RGWB or red, green, white, and blue layout seen in many recent OLED panels. Strictly speaking, pure RGB and without the white sub pixel LG adds to boost brightness offers both better rendering precision and superior colour fidelity. For the record, Asus has announced a monitor with LG's new Tandem OLED panel that does have pure RGB-stripe subpixels, the ROG Swift OLED PG32UCWM. But that monitor is not this monitor, and the Swift models always command a much higher price.

Whatever, as is usual for all these OLED panels, response is rated at a startlingly scant 0.03 ms. In short, pixel response is essentially a solved problem with modern OLED panel tech. Rounding out the main features are dual HDMI 2.1 inputs, DisplayPort 1.4, a USB hub and support for G-Sync adaptive refresh technology on Nvidia GPUs and FreeSync on AMD GPUs.

Incidentally, this may be an Asus Strix model and therefore not its most premium offering, but the build quality is still decent. Okay, you miss out on pivoting the panel into portrait mode. But all the other adjustments, including tilt, height and swivel are included and, frankly, it's every bit as well put together as an 8K Asus monitor I happen to have in at the moment that sells for an incredible $8,800.

With all those details covered off, well, it's time to get to the main event. How good does this new LG Tandem OLED panel tech look in 32-inch 4K format and, likewise, how does it stack up against the Samsung QD-OLED competition as represented by the MSI MPG 322UR X24?

Out of the box, it's immediately apparent that this new panel has both more punch and more colour vibrancy that the previous LG-sourced 32-inch 4K OLED panel. But there are caveats.

Subjectively, the full screen brightness is a little disappointing. With a large proportion of the panel being driven hard, it's definitely duller looking, subjectively, than the MSI. If that seems to contradict the paper specs, well, this Asus monitor has to rely on LG's additional white subpixel to hit that 330 nits full-screen figure.

In practice, the white subpixel doesn't automatically boost brightness for colours as opposed to pure white. So, much of the time, that 330 nits rating is a bit notional. As with most OLEDs, you can boost the brightness by disabling the ABL or automatic brightness limiter, which Asus refers to as the Uniform Brightness setting.

However, more than with QD-OLED panels, the overall panel brightness varies substantially depending on how much of the panel is being lit up. It only requires a disappointingly small application window, such as a web browser, showing a predominantly white image, to have the panel dimming quite noticeably.

Even using dark mode for most apps, the panel will quite often dim in a manner that the QD-OLED competition doesn't, not quite. That said, this monitor can still deliver some spectacular visuals for HDR content.

It's at its best when rendering small bright details on a dark background. That's when you'll get the full 1,500 nits and a truly eye-popping experience. That's not the reality for most games, most of the time of course. And for day-to-day gaming I'd give the subjective edge to the MSI monitor and its Samsung panel. It's that little bit more vibrant, that little bit more immersive.

Of course one thing the MSI lacks is a 1080p dual mode. So, how does that factor into the equation? Certainly, this Asus Strix is lightning fast running at 480 Hz. The motion clarity is pretty stupendous ripping your mouse around in an online shooter like Counter-Strike 2, ditto the lack of latency.

The image quality at 1080p is also among the best I've seen in terms of pixel definition on a 32-inch 4K panel running in dual mode. But the thing is, if you're really serious about esports, do you really want to be playing on a 32-inch monitor?

I suppose you could mimic the apparent size of the preferred 24-inch monitor by pushing this thing further away. But I'm not totally convinced that's a satisfactory solution for someone who plays loads of online shooters. And if you don't, I can't see you using the dual mode much.

What I can say for sure is that I still give the slight edge to Samsung's QD-OLED tech. Yes, even the latest QD-OLEDs have a slight purple tint in bright ambient light that you entirely sidestep with LG's Tandem OLED panels.

But, overall, QD-OLED still feels noticeably more vibrant and colour rich. Moreover, the variable brightness on this Asus's LG panel when you disable uniform brightness in order to get the best performance is also more noticeable and more distracting.

All that said, there's a price differential to take into account. This Asus goes for $899 to the MSI's $1,099. Personally, I think the MSI is worth the extra $200 for a bit more subjective pop and USB-C connectivity. But the value proposition is certainly very tight between the two. And if you don't care about USB-C but would like the 1080p 480 Hz dual-mode feature, well, the Asus ROG Strix OLED XG32UQWMS is pretty compelling.

Is this the monitor cashed-up PC gamers have been waiting for? At last, a 39-inch ultrawide with decent pixel density thanks to its 5K2K resolution, plus the latest LG OLED panel knowhow? The new LG UltraGear Evo GX9 39GX950B certainly sounds pretty sweet, even if it also costs a pretty penny at $1,799 or £1,549.

The sordid matter of money aside, there's just one snag. The Ultragear 39GX950B does indeed offer LG's hot new Tandem OLED technology. But there's another otherwise almost identical 39-inch 5K2K panel from LG Display that has one notable advantage over this monitor, the inclusion of true RGB-stripe subpixel structure.

It turns out Alienware has announced a monitor using that LG panel, the Alienware AW3926QW, and, to be frank, it looms large over this new model from LG itself. However good it is, you'd really want to see how that Alienware shapes up before pulling the trigger.

Still, the LG Ultragear Evo 39GX950B does look epic on paper. We're talking 5,120 by 2,160 pixels across that 39-inch diagonal. To provide a little context and to define this panel in terms of pixel density and vertical screen size, this is essentially a 32-inch 4K 16:9 monitor taken out to 21:9 aspect and 39 inches.

For many fans of high-end gaming displays, the basic proposition here is just about perfect. The 143 DPI pixel density is enough for crisp fonts and lots of space for day-to-day computing and multitasking while also delivering oodles of in-game image detail.

At the same time, the ultrawide aspect and huge 39-inch panel size makes for epic immersion. Then add in the latest LG Tandem OLED panel tech and, well, you get the idea. This thing promises to be a bit special.

Beyond the basic form factor, of course, there are plenty of other details to absorb. In terms of brightness, ever a key OLED metric, LG claims 335 nits full screen and 1,500 peak HDR in a small window. Those figures are as good as PC OLED monitors currently get.

Speaking of HDR, the 39GX950B has VESA DisplayHDR True Black 500 certification, which is a hallmark of the latest LG Tandem OLED panels. Response comes in at 0.03 ms, which is standard for pretty much all current OLED panels and miles faster than any LCD-based monitor.

As for refresh, at the 5K2K native the 39GX950B tops out at 165 Hz. However, it supports various dual-mode alternative resolutions, the fastest of which can achieve 330 Hz. That's fairly nippy by any sensible measure, though it's pretty hard to imagine anyone really serious about esports and online shooters going for a huge ultrawide like this.

Having said that, the 39GX950B does offer the option of running at 1080p and 330 Hz in a 24.5-inch window, which pretty neatly mimics the preferred form factor for esports. The image quality is pretty clunky and pixellated, as it is in all of the dual modes. And I'm not totally convinced that many esports fans will actually want to play like that.

It certainly looks pretty odd to have a small 24.5-inch section of the huge panel active and rest fully blacked out. But there's no question it works and means this monitor is a thoroughly versatile gaming tool.

If the downscaling support is comprehensive with this panel, so too is the upscaling. Obviously, some kind of upscaling is now pretty much standard with any remotely recent graphics card, in the form of Nvidia's DLSS, AMD's FSR or even Intel's XeSS. But LG has included its own "AI" upscaling functionality with the 39GX950B via an integrated image processor. But hold that thought, we'll come back to how effective it is and how it might compare with GPU scaling in a moment.

Rounding out the speeds and feeds, you get HDMI 2.1, DisplayPort 2.1 and USB-C connectivity, the latter with 90 W of power delivery for single-cable connectivity to a laptop, albeit a gaming laptop will typically need rather more than that.

Last and certainly not least on the specs, the panel has a matte anti-glare coating rather than glossy and a mercifully gentle 1500R curve. Most people know if they are glossy or matte fans, but suffice to say that the heavy majority of current OLED gaming monitors are glossy.

One further detail involves the warranty. Where most big monitor brands offer three-year cover, including burn-in, LG is only offering two years. That's disappointing on several levels. For starters, LG Display, the subsidiary of the sprawling LG empire that makes the actual OLED panel (as opposed to LG Electronics, which makes this monitor), says its new Tandem OLED panel tech is more durable than previous generations. So, why not back it with a three-year warranty?

Moreover, given the hefty asking price of around $1,799 or £1,549, the thought of being on your own when it comes to burn-in after just two years is pretty off putting. That alone makes this monitor tricky to recommend.

But let's park that caveat for a moment and consider this monitor's performance. What's it actually like? Well, it's definitely punchier than monitors based on LG's previous WOLED panel generations. Oddly, it's also a more consistent performer than the Asus ROG Strix XG32UQWMS I reviewed recently.

I say "odd" because that Asus uses essentially a 32-inch 4K version of the same LG Tandem OLED panel tech. But I found the brightness variability on the Asus to be awfully distracting if you disabled the uniform brightness feature in order to maximise performance.

The Ultragear 39GX950B does still suffer from that issue, and probably more than comparable OLED monitors based on Samsung QD-OLED tech. But not enough for it to be a major flaw. Overall, the full-screen brightness is pretty decent and this panel can actually be too bright if, for instance, you tend to run most applications in dark mode and then have a third of the screen lit up by a bright browser window.

Of course, that's a good problem to have. Turning brightness down is simple. There's not much you can do, on the other hand, if max brightness isn't good enough.

So, as a desktop general computing panel running in SDR mode, this LG is pretty sweet. Fonts are sharp and crisp, there's loads of desktop space for multitasking, and the panel is plenty bright enough. If you really want to split hairs, I think the latest Samsung QD-OLED panels still look a touch more vibrant and saturated. But it's not a huge delta.

For HDR content, meanwhile, it's likewise mostly good news. LG's OSD menu is a bit baffling, but once you've got everything properly set up, the results can be spectacular, especially on a panel this size.

As ever with an OLED panel, it's at its best showing small bright details on a dark background. Neon lights in a nightclub scene, fireworks in the night sky, a large flame burning in the inky darkness, that kind of thing. Then the full impact of the 1,500 nit peak HDR brightness really hits home.

It's a touch less impressive for outdoor daytime video and game scenes, where the full-screen brightness limitations can leave it looking just a tiny bit underwhelming. To be clear, it's far better than earlier LG OLED panel generations in that regard. But, again, my preference for that kind of rendering remains Samsung QD-OLED.

That said, it's an open question just how good LG's new RGB-stripe panel as used by the aforementioned Alienware will be by comparison. In theory, and because it doesn't rely on an additional white subpixel to boost brightness, that RGB-stripe panel should have superior colour saturation and look more vibrant. Exactly how that pans out, well, we'll have to wait and see.

While we're talking HDR visuals, the SDR calibration in the best HDR mode is pretty wonky to the point where you'd never leave it in that mode, day-to-day. You can actually get decent SDR visuals in one of the other HDR modes. But then that compromises HDR performance. Long story short, you're going to have to switch monitor modes depending on the kind of content you are viewing, which is a bit of an annoyance.

As for the AI upscaling feature, it's awfully hit and miss. For instance, it doesn't work at all at 3,440 by 1,440, which you'd think would be the ideal alternative setting to 5,120 by 2,160 native in terms of knocking the res down to release some performance.

At other resolutions like 1080p, meanwhile, it only works at certain refresh rates. Once enabled, yes it does make things a bit crisper, but it's really not comparable to DLSS or FSR. In short, in its current form the built-in AI upscaling feels pretty underbaked.

Still, if you add panel speed to this monitor's other qualities, the overall experience is frankly epic. It's so huge and so super sharp and so immersive and just so awesome. And yet I can't help having a few reservations.

First, I'd love to see the latest Samsung QD-OLED tech applied to a panel this size and resolution. I'd likewise like to see this monitor with a glossy coating. That could be fantastic in terms of heightening the sense of visual pop and contrast.

And, again, I'd really want to see that upcoming Alienware with its LG Tandem OLED RGB-stripe panel before pulling the trigger, especially as the Alienware looks like it could be as much as $600 cheaper.

Then there's the disappointing two-year warranty. It really is hard to look past that at this price point. All of which leaves me in a state of cognitive dissonance with this monitor. In some ways, it's the best gaming monitor I've ever seen. And yet it's ultimately impossible to issue a strong buy recommendation. At the very least, anyone considering the LG UltraGear Evo GX9 39GX950B really has to wait and see how that cheaper RGB-stirpe Alienware turns out.

The Secretlab Atlas has a lot going for it. It fits snugly into a home office and it's very comfortable for use throughout the day (with regular breaks, of course). It's even good for a spot of gaming in the evening, if you're so inclined, providing you understand that this isn't explicitly a gaming chair. It's a task chair—whatever that means.

Google tells me a task chair is a chair designed for sitting at a desk. One website described a task chair as a chair specifically designed for doing tasks. Fascinating. Nevertheless, Secretlab is using the designation to separate the Atlas from the ever-popular Titan Evo, and in doing so, alluding to its sleeker form and more civilian appearance.

The Atlas looks excellent in situ. It's a smart-looking chair for the discerning gamer. It's halfway to a gaming throne and definitely passable in a modern office without raising eyebrows. Doubly so providing you pay for the premium model. Though this is where things get tricky for the Atlas.

The Atlas is on the expensive side for the features it offers, especially when compared to the Titan Evo. It's not easy to beat the best gaming chair of the past four years, and while the Atlas does serve a slightly different audience, for us gamers, the Titan Evo is yet to meet its match. It's not a whole lot more expensive than the Atlas—at times, I've even seen it going cheaper.

Features

There are a range of options available for the Secretlab Atlas, from basic and premium finishes to a range of colours and materials. I'm testing the Atlas Dune+ model, in a size L. That means it features two of Secretlab's proprietary materials: NanoFoam Composite and SoftWeave Plus fabric. The NanoFoam being softer than cold-cure foam used on the basic model.

The premium version of the Atlas comes with colour-matched parts, from the magnetic head pillow down to the armrests, beneath the seat, wheelbase, and the castors. This is made possible through the use of a reinforced nylon construction, as opposed to the aluminium alloy found on the Titan Evo. My gut tells me that metal is best but I couldn't point to any research to back that up. Just a hunch. The cheaper version is black except for the backrest.

The Secretlab Atlas is noticeably slimmer than the Titan Evo, Noblechairs Hero, and ThunderX3 Solo 360 that my colleagues are sat in close to my desk at PC Gamer Towers. It much more closely resembles the Fractal Design Refine used by our James, in terms of both the backrest and its more professional aesthetic, but the Atlas is a little more reserved. Unlike the Refine, however, the Secretlab has no adjustable lumbar support.

Secretlab claims the fixed backrest, called RE-CURVE, is "sculpted to support your entire back". I find myself agreeing with that statement after a couple of months with the chair as my daily rollabout. That said, it does lack the flexibility that other chairs offer, including the Titan Evo, which offers an excellent 4-way lumbar system called L-ADAPT.

There are a number of possible adjustments on the Atlas. Besides the usual height adjustment—the seat is lower to the floor on the Atlas compared to the Titan Evo—the seat also slides forward and backward on a horizontal axis. I recently spoke with a physician about how to sit (and how I was doing it wrong) and they told me you should be able to get two or three fingers between the seat and the back of your knees for an ideal setup. Secretlab says that's about 5 cm/2-inches in its promotional materials.

The Atlas will recline up to 120° once unlocked, using a lever with a handy visual indicator to show when the leaning mechanism is locked or unlocked. There are four levels to the tilt force, adjusted using a twisting selector on the chair height stalk. These levels range from leaning back with little to no resistance to having to actively push back on the chair to get it to recline even partway. I like to use a higher setting to bob in place a bit—hey, I'm just a bit wild like that. I'm one errant purchase away from a rocking chair, I swear to God.

The armrests offer height adjustment, angle adjustment, and front to back adjustment. That makes them 3D, to the Titan Evo's 4D ones. They will move forwards and backwards when only a moderate amount of pressure is applied, which is a bit of a pain for someone that shifts my chair around a lot while switching between sitting and standing, but otherwise, they're good enough.

The top of the armrest is comfortable for long-term use. Though it does not feature the same CloudSwap Replacement System as the Titan Evo, which means you can't switch these out as easily once worn.

At least my favourite feature from the Titan Evo has made it over to the Atlas. The magnetic head pillow. This can be easily removed, replaced, and moved around without messing with straps. The Atlas actually improves on the design, too, with a more rounded shape that better fits my noggin.

Ergonomics

So far I've very much enjoyed my time with the Atlas. It's a firmer seat than many I've tested, but that's no bad thing. I've been putting it to the test in a range of seating positions: legs crossed, laying back, fully reclined, one leg under the other… you name it. It deals with the cornucopia of weird human sitting positions very well. After a frank meeting with that physician about how I sit, I'm trying to change things up more regularly throughout the day, and the Atlas allows for that variation.

My main points of comparison for the chair are the Herman Miller Embody (Logitech G), which I've used every day for roughly four years; the Razer Fujin Pro, which I've been using in PC Gamer Towers for a couple of years; and, of course, the Secretlab Titan Evo, which I used in the office for years and I still sit in every time I shoot a video for the YouTube channel. We must have eight or so of those around.

First off, the Atlas easily beats out the Razer Fujin Pro. I'm yet to find a mesh chair that I enjoy to rest on as much as any with foam. I used to sit in a Cooler Master mesh chair and felt about that much the same as I do the Fujin Pro: foam is just a little too stiff. Over the course of a day's work, I'm aware of the pressure. That's not true of the Atlas, nor is it the Herman Miller, to be fair.

The Herman Miller is lovely to sit in. Though the backrest is stiffer on the Atlas, which I currently prefer, and I'm finding myself making more use of the recline on the Atlas. That is likely down to breaking out of old habits with a new chair. Though perhaps because I find the recline a touch more supportive on the Atlas. I definitely prefer a stiffer backrest over the flexible one on the Embody.

Looking around, there are anecdotal reports of larger users finding this chair uncomfortable. At 6'1" and 80 kg, I haven't run into any issues, but my thighs and back do sit close to the edge of the part of the backrest and seat where they shoot off at an angle—the racing seat sorta bit. Spare a thought for the sizing as even Secretlab notes the Atlas as more "form-fitted" to the Titan Evo's "spacious" fit.

I'm sitting very comfortably in the Atlas. It's in the Goldilocks zone for me, with firm yet comfortable materials, a stiff backrest, bouncy recline, and squishy head cushion. I'm not in any rush to swap back to the Embody.

Aesthetics

I had feared for the feng shui of my home office by dropping a sand-coloured gaming chair smack-bang in the middle of it. But, to my complete surprise, it makes my office look larger. And it's not just me. My partner walked into the room after I built it, and without me mentioning a thing, said the chair makes the room feel bigger. Weird.

Hey, I'll take it. I think it's some sort of optical illusion with the bright chair, my dark grey walls, and the slim backrest. The chair does look pretty sweet in situ, too. There are other colours to choose from, including grey, black, and white/black leatherette. I've seen pretty much all of them in-person and I'd be happy with any, except perhaps the cheaper two-tone version without colour-matched parts.

Value

The Atlas is a cheaper alternative to the Titan Evo in the US, with prices starting at $499. That's for the basic model in a regular size, specifically the black leatherette model. That is the only basic model I'd actually purchase myself, anyways, as the non-colour matched parts incidentally match the backrest. The Dune SoftWeave Plus model starts at $519.

The premium models look much nicer and the only ones I can recommend on comfort, as they are the only ones I've sat in long-term. These start at $699. The large model I'm using is $719. That's not a low cost gaming chair, by any means, and the Atlas has stiff competition at this price. Not the least bit from Secretlab's own Titan Evo, which starts at $579. The premium NanoGen model costs $799, and a large will set you back $849. That's a $130 difference—not a trivial cost to upgrade from the Atlas to the Titan Evo, like-for-like. But maybe you don't need to go like-for-like?

If I'm honest with myself about what I'd personally buy of the two, it'd be the Titan Evo. And not the NanoGen version. I would happily take the non-NanoGen Titan Evo for $629, which is $90 cheaper than the Atlas. Having sat on the standard leatherette Titan Evo for years in the office without issue, I'd be content with that. It's more fully-featured than the Atlas with adjustable lumbar support, replaceable armrest cushions, and an aluminium wheelbase.

The Atlas does beat the Titan Evo in two key respects: style and size. I won't discount the impact these have on purchasing decisions, as I've already had a friend tell me they're considering buying the Atlas due to its shape alone.

Looking further afield, the Fractal Design Refine is not a chair we rated particularly high but it is a similar sorta thing with an adjustable backrest. That's $600/£475. I think I'd stick with the Titan Evo. Though there is a budget alternative, and that's the ThunderX3 Solo 360. This offers asynchronous tilt, adjustable 3D armrests, decent fabric, and it's half the price of the cheapest Atlas. Except the Atlas doesn't feel like a budget option. The Solo 360 still does: it doesn't include lumbar support or a head pillow.

For larger bodies, we like the AndaSeat Kaiser 4 XL. It's extremely cushioned and well-provisioned for lumbar support. For $579, it offers a lot compared to the large Atlas, though it loses out on that sleek appearance. It is huge.

So, the Secretlab Atlas does offer a high level of quality and attention to detail that see it clearly rise above many budget offerings. Though, there's been a lot of cuts here compared to the Titan Evo, such as the wheelbase, armrests, recline, and lumbar support.

The Titan Evo often drops to around $529 during yearly sales events, like it did back in May. As I write this, in my native land of Britain, the Titan Evo is cheaper (£419) than the Atlas (£429) in the company's Summer Sale. The Atlas is understandably yet to be discounted, but that's a tough act to follow.

The premium models, especially in the large size, are really pushing the limits of what I'd pay for the Atlas. That's a shame because that's what I'm sitting on and I really like it. The Atlas Dune+ is comfortable, stylish, and I'm happy enough to sit and game in it all-day. It's just the Titan Evo is tough to beat, even the cheaper standard model.

"A nocturnal farm sim where you're a vampire" is one of the better hooks I've heard lately in a subgenre that's long become overstuffed with too many same-y, saccharine rural community sims. Moonlight Peaks' slightly spooky theme is a welcome twist, but that's where its originality ends, because it struggles to ever turn that concept into a story with any bite.

You play as Dracula's spawn, storming out of Dad's castle to take up farming "cruelcumbers" and "blood tomatoes" at the overgrown family retreat in small town Moonlight Peaks, where families of vampires, witches, seers, and werewolves are mired in years of bickering. There's a whole lot of all-caps dialogue in this town.

Eternally familiar farming

It's got the requisite activities checklist for a post-Stardew, Animal Crossing-like farm sim: crops, animals, foraging, fishing, bug catching, a museum, cooking, friendship and dating, decorating, and a bit of crafting. All those basics are executed competently enough that I can shut my brain off and have a good time for a few hours. Digging tidy rows for my blueberry plants, using kegs to make wine, and running off to mine copper for an axe upgrade is all comfortingly familiar, but as soon as I try to look any closer the cracks show.

Unlike more open-ended cousins like Stardew Valley, progression in Moonlight Peaks is extremely main quest-driven. There's no leveling your farming skills or unlocking buffs here. It's all about making money to fund the string of crafting and collecting tasks that help revitalize the town and repair the relationships between its supernatural families. Despite that, I often go days at a time without an active quest to pursue, going to bed early after watering my crops and talking to a couple villagers as I wait for my next goal to trigger.

In between quests and farm planning there are several side activity minigames like flower arranging and pottery, which are genuinely neat, though there's not much incentive to engage with them after making a couple custom decorations for my farm. Some are used as date activities, which is a cute concept, but each character only goes on one date type, so players who choose to only woo one character will be stuck endlessly embroidering or marshmallow roasting their way to marriage, which gets old fast.

For those who do date around, being able to play the field before committing to marriage is nice, and common now in farm sims. With just a few families living in Moonlight Peaks, though, I've found myself simultaneously dating parents and their adult kids, or all four members of a witch coven, or two siblings, which feels a bit odd.

The relationships panel does track gifts that I've given to the locals to earn their friendship, but only the past eight items, duplicates included, which isn't nearly as useful as a discrete list would be.

Every interesting idea or variation on farm sim standards that I can find to praise in Moonlight Peaks comes with an asterisk, which is most true of its story and characters.

Totally de-fanged

Thanks to that driving main quest, there's a ton of story content in Moonlight Peaks. Each character has four cutscenes doled out as we become friends, with more available if I choose to date and marry them. Many of the main quests also involve a handful of the cast getting together to banter and bicker. It's a total feast of character interaction compared to so many other farm sims out there. But yet again, a caveat: The writing is offensively bland.

I don't expect literary innovation from a genre whose current trends prescribe a cast of character tropes. There must always be a friendly rancher, fastidious mayor, plucky carpenter, and a charming bartender (etc.), most of whom should be hot, single, and down to date. Moonlight Peaks has done more than just cleave to character archetypes, though. It's managed to not find a single original string of words in that bounty of character dialogue.

The coven's relentless flirt of a warlock Noel is all confident bluster, which of course begs a relationship arc to uncover his inner sincerity. In one of his early cutscenes he says "there's more to me than this front," in nearly those exact words. "When I'm around you… I don't have to pretend," Noel later says in the scene confessing his feelings to me after we've been dating. Every character interaction is full of these same trite lines I rolled my eyes at in network TV shows 20 years ago.

The tone is inconsistent. Interactions with alcoholic vampire dad Orlock are played for laughs at every opportunity, but his behavior is treated as a serious emotional burden for his adult kids in their own character scenes. Then it's right back to blacking out gags. Its characters all fail to find a compelling identity with inconsistencies like Elvira expressing excitement over seeing pumpkins in the autumn festival but being totally disinterested in receiving one as a gift.

I just haven't managed to care about the story when Moonlight Peaks can't decide if it wants to be silly or sincere, sweet or dark. It's possible to juggle those moods at different moments, but this cast consistently drops the ball.

Moonlight Peaks works as a second screen game for farm sim fans who just want a new aesthetic in which to decorate a home and lay out perfectly even rows of tomato plants. This subgenre is completely packed with options at competitive prices, though, so I can't eagerly recommend one that feels so thin.

As I said in my Razer Kiyo V2 X review, the 1440p webcam market is almost as niche as the 4K one, and that's almost entirely because of its price point. With the likes of the Emeet Pixy regularly sitting at the $90-100 mark, the $100 Kiyo V2 X really struggles to justify itself, and yet Dell's Pro 5 2K webcam is $20 more than that.

But Dell does have a few tricks up its sleeve to justify that extra cost, even if I think it's comparably as attractive as the Kiyo V2 X. One big feature, Windows Hello, is genuinely good and shows that resolution is important, but it's not everything when it comes to picking out a webcam.

Dell have crammed a Sony sensor in its 2K webcam and, for the most part, it's rather solid. Texture in my hair is captured well, lighting off the side of the webcam doesn't look too blown out, and autofocus is fairly snappy. Its focus doesn't match the near-unnatural speed of the likes of the Obsbot Tiny 3 or Tiny 2 Lite, but it's still more than good enough for showing off a box or label to your buddies over Discord.

One area I do think it drags behind other webcams I've tried is in colour accuracy. Video and pictures just look a little lighter than real life. For instance, the red streak in my hair comes through well in the DSLR-like Yolocam S3 but almost looks orange in the Dell webcam. The same goes for colours on my shirt and my background. This is definitely a matter of preference, but for me, I wanted to immediately tone that down.

Unfortunately, though Dell's own software is fairly easy to use, I couldn't quite tone down the image as I would have liked. Bringing up vibrance makes colours look a little more natural, and sharpening contrast helps it along, but it just looks like a brightness filter has been placed atop the image, which I can't seem to take off.

As well as this, I'd say none of the presets are worth using in a moderately well-lit environment. The 'smooth' preset cuts out all sharpness and looks generally too glossy, and both the 'Vibrant' and 'Warm' presets are simply too bright. The 'default' preset offers a decent enough level of quality, but one that feels like it has more of a software helping than I would like.

One could argue that treatment is necessary on an otherwise cheaper webcam to try and pull up that quality, but 1440p isn't exactly low. In fact, many bits of software won't even run at 4K if you want them to, so 1440p should be a great middle ground between quality, accessibility, and budget. The Sony sensor inside Dell's webcam shouldn't need that much of a software hand to help it out.

As far as design is concerned, the Dell Pro 5 is inconsistent. The webcam itself is chunky and sturdy, with the magnetic stand offering a comfortingly heavy weight, but the stand connects to the webcam with a magnetic pull, and that pull is simply not good enough. It won't fall down by itself, but if you accidentally tug on the wire a little, the webcam will slide, messing up your otherwise perfect framing.

This means that, despite the cable being one and a half meters long, you actually won't want to use it any more than 1.4 meters away from your computer; you need some slack in the wire to stop it wiggling on the stand. I suspect a chunkier wire would cause all sorts of problems, as it does for the weak magnetic stand of the Obsbot Meet 2, but the Dell webcam has a single attached wire. I don't like this, as it adds another point of failure to the device's design. If the wire goes, the webcam goes with it.

I do like the webcam's built-in privacy filter on the top. That's a welcome feature, and it's easy to use, but once again, that magnetic stand means that using the filter moves the webcam. This isn't a huge problem as you can simply adjust the webcam whilst you're flicking the switch, but it's a sign of just how weak the magnetic connection is.

The webcam's built-in omni-directional microphone picks up sound decently, with passable quality. Recording sounds a tad hollow, but it doesn't have any awkward filtering on it, and is generally fine for a quick call or conference.

The Dell Pro 5 Webcam feels like it's designed for more of an office setting, with its smooth, brightened software enhancement, ease of use, and, of course, its Windows Hello functionality. With this, you can register your face and sign into Windows by simply sitting in front of your computer. It works really well in practice and is a very strong feature for an otherwise just okay bit of hardware.

It's a shame that the lack of IR sensors means other webcams can't do the same, as I like the feature a lot, but I certainly wouldn't trade out my day-to-day webcam for the Dell one just because of Windows Hello.

Ultimately, the Dell Pro 5's biggest weakness is its price point. Though I do prefer it to the Kiyo V2 X's 1440p capture, it is very expensive for visuals that are worse than some 4K offerings and lacks the level of customisation necessary to really fine-tune it. The tethered cable and weak stand only make it a worse all-round package.

Ultimately, when you can get 4K webcams on a gimbal for smart tracking that provide better quality video for less money, it's Dell's name recognition, ease of use, and Windows Hello functionality that might pull it ahead. And for me, it's only that final quality which would grab me.

My train spins through the air after launching off a skyscraper-high vertical ramp, swiftly twisting onto another bolt-upright bit of track, my view smothered with speed lines, Japanese onomatopoeia, and sparks as enthusiastic allies yell encouragement. A gorgeous vista stretches out below me, Japan's countryside filled with cherry blossom trees and bright blue water.

In these brief and highly curated moments Denshattack appears to be an exciting combination of Jet Set Radio's cel shaded style and Tony Hawk's tricks and challenges, with an added dash of Taito's arcade classic Densha de Go for good measure. It's an impression that sadly soon goes off the rails, any interaction longer than a social media clip's worth of action revealing a fully loaded freight train's worth of issues.

I may technically be controlling a train, but bar a token nod to not derailing on curves the metal box I'm in charge of never behaves anything like one, even if I temporarily set the acrobatics aside. There's no inertia, no weight pulling me back on steep slopes, no assistance from gravity on the way down a ramp, no careful adjusting of a virtual mascon to smoothly slow down for a corner or push for flat-out bursts of energy on a long straight.

It could be anything. A person. A dirt bike. Shadow the Hedgehog. Tracks themselves are broken into tiny chunks and discarded as often as the game's whims deem necessary, replaced with anything from half pipes to trampolines and ferris wheels.

I never expected Denshattack to play anything like a serious sim, but I did hope the train aspect of the game might involve more than letting me pick and decorate a few train-like shapes, especially as the (Japanese) word is literally in the title.

Ignoring the game's self-chosen focal point entirely and pretending it's purely a futuristic trick-based score attack experience does little to improve things. There's an inescapable feeling while I play that it would be happier if I didn't get in the way. Personal creativity is constantly pushed to the side, unwelcome in spite of all the combo multipliers and flips I can do. Every stage wants to tell a story at all times, to be a wild string of events where pure-hearted people shout nice supportive things as structures crumble and wacky scenarios pop up around them: A giant mech. A laser-spewing kaiju. A kabuki play. There are even playable Rock Band and Ikaruga homages to work through.

I'm rarely trusted to make my own fun. There's always some weird quirk to work around (at the most basic, optional rainbow tracks that are completely invisible, and therefore, unable to be spotted and planned for, until a gauge fills up). I can't strategise or create combos on my own terms.

Or guarantee I'll react in time to whatever "fun" eye-catching stunt is coming up either. The game may haphazardly offer icons warning of upcoming dangers, but when a red "!" can mean anything from "obstacle on the line" to "a giant shadow monster is about to rise out of the sea" and I've got half a second at most to react, they're of little practical use. Occasional poor colour choices only exacerbate these issues, long strings of orange lanterns hanging right beside orange grind rails, lines of yellow lights near yellow track, green air currents set against a sea green background. I'm constantly fighting the stages themselves. Failure does little beyond setting me a short way back so I can try again, the rote repetition only leading to increasing irritation instead of a meaningful opportunity to learn.

I either do what I'm told when I'm told to do it or… Denshattack has no "or," too much of the game playing like a string of QTE-like gimmicks. Me being in complete control is never the game's priority.

Attempting to make do and master what is in here is routinely thwarted by the game's wobbly sense of internal consistency, my capabilities and even plain old physics only intermittently available. My ability to race at odd angles and perform Titanfall 2-like wall runs flickers on and off like a faulty lightbulb—that bit of track, that vertical surface—and attempts to think outside the box and perform these stunts outside specific moments is brought to a crashing halt, Denshattack tut-tutting about my balance and deciding any contact with smooth perpendicular walls is a punishable offence. The water's surface must never be touched except in the occasional stages where I'm allowed to ride underneath it for extended periods of time. Crashing into things is an instant fail unless the game decides that actually crashing is OK for this bit or this particular object.

It doesn't feel fair, and the lack of clarity makes it difficult to improve my own skills when what is and isn't allowed varies so drastically from one section of a stage to the next.

Denshattack doesn't seem to know what it wants to be. For a game with such strong anti-corporation/environmentalist themes—freeing Nara's famous deer is a stage-long event—it's a little odd to see so many optional challenges involve some variation of "scare the animals by honking the train's horn." Trains are surprisingly irrelevant to a story full of them. Tricking struggles to rise beyond the level of a simple distraction: Abilities gained throughout the story don't broaden my toolset or scoring capabilities, and retrying an earlier track completely disables any later abilities gained. Tricks are just a way to survive the latest set of scripted setpieces.

It isn't clear and consistent enough for serious score attack play, yet it's also far too fussy to allow the raw absurdity of moments like controlling a shark or countering giant baseballs with aerial tricks to sweep me up in a wave of comedic positivity. This train has gotten lost somewhere between two diverging tracks.

Don't scoff at it being a DLC: Revelations is the best Doom has ever been. Hell, it would be worthy of being a standalone sequel if it weren't so tightly woven with the base game in ways id Software joyfully refuses to do with its full-fledged follow-ups. While I adored The Dark Ages' visceral demon-crushing, its level design fell flat. Some levels were too linear; some were needlessly open. So I wasn't expecting to enjoy the DLC too much before I went in.

Turns out, Revelations is a perfect blend of Eternal's movement and level design, and The Dark Ages' timings-based combat. And you know what? I think it's the most fun I've ever had in Doom—and it's almost the size of a full game, clocking in at about 10-15 hours. A damn shame considering Xbox laid off id Software developers just days before its release. Talk about the dark ages.

Tip of the spear

Only a short way into Revelations, Doom Guy loses all his gear, including his Shield Saw—but don't worry, it's getting replaced with something even better. Enter the Chain Spear. This is the big gimmick of the DLC, trading the defensive power of the shield for massively increased mobility. And in typical id fashion, this new weapon is a multifaceted tool that's so creative that it could support an entire game on its lonesome.

You can attach it to enemies to launch yourself around, even upgrading it to let you orbit them while you blast away. You can slash enemies, destroying metal and shields. This can also parry green Hell Surge attacks if you time it right, similar to the Shield Saw in the base game. The bricks tied to Doom Guy's feet have been untethered now too so you can dash in any direction.

The spear is just as gnarly and badass as you'd expect, but it's home to a remarkable amount of depth, especially when you start unlocking the various upgrades. Dodging and parrying attacks earns you Empowerment, which you can use to throw a stunning spear (great for flying and spectre enemies), stab it into enemies to superheat metal and cause explosions, or, my favourite, groundpound to cause enemies to drop health.

You'll regain access to your old weapons and, soon enough, the repaired Shield Saw, opening the door to yet more combos. Revelations encourages you to master the flow of combat more than ever with tougher demon encounters and more tools to play with.

It's a hell of a lot of buttons to juggle at any given time, but it's one of the most rewarding, satisfying combat systems in the genre.

Level up

While your mileage with the combat still largely hinges on how much you like the gimmick of countering specific attacks, the level design is a clear step up in every regard. For one, I was consistently amazed by how cool Hell, Purgatory, and the Lovecraftian-inspired Cosmic Realm look. Hands and faces protruding from medieval castle walls. Cosmic beasts in unfathomable, shifting labyrinths. It's rad.

The new missions are among Doom's greatest. They're intricate and layered, striking a great balance between mandatory encounters, optional puzzles, and endgame clean-up. The boring binary choice of going one direction to progress or the other to find the obvious secret path is, thankfully, gone. Though that does mean backtracking is common if you're not stalking the map screen, and you can't return to levels until after you've beaten the story. It's pretty punishing if you miss some Platinum on your first run through. Given the focus on combat mastery and perfect flow, having to retread old encounters isn't a bore, at least.

I was pretty surprised by how good most of the puzzles are. You're still moving blocks and finding keys, but there are more layers and some of them actually had me stumped for a minute or two. It's the polar opposite of the base game, which felt like it begrudgingly included them out of an obligation to its predecessors.

I loved returning to the main puzzle box hub area after each mission to see what else I could access, such as lowering the water level or pulling down obelisks for treasure. This area contains a lot of the endgame challenges as well, which you'll unlock after beating the story, like time trial combat challenges, silly levels inspired by the original Doom, and tough bosses.

The rewards from exploring are also much improved, harkening back to the satisfaction of checking every nook and cranny in Doom 2016. Frankly, I got fed up with most of the secrets being barely hidden, boring Life Sigils and Codex pages in The Dark Ages' main campaign. Here, you're assembling relics for Platinum to grab those meaningful upgrades and getting new skins and toys on a consistent basis. Revelations realises how fun it is to navigate and comb over well-crafted Doom levels.

Revelations is, on the surface, just more Doom. You shoot a hell of a lot of demons in a boxed-off room. But the new and existing tools at your disposal are perfectly intertwined, while the smart level design elevates the combat and exploration. Even the story is great here, something I usually don't care for in Doom—I love the retro stylings of the flashback levels, and how being a prequel lets it retroactively tie into its predecessors in cool ways.

It's just a crying shame that such a refined, full-throttle Doom experience is not only relegated to a DLC for the most controversial entry in the series, but that the team that made it has lost talent thanks to Microsoft's studio-crushing hammer.

When Qualcomm released its first batch of ARM-based Snapdragon X processors two years ago, they worked well enough, but failed to set the PC gaming world on fire. It's not really what they were meant for, and if decades of buying computer parts have taught me one thing, it's never to buy into the first generation of a new product. I still do it, of course, as the Oculus Rift sitting unloved in its box shows, but I'm always left with the feeling that I should know better.

The main downside of the OG Snapdragon, once software compatibility issues were smoothed out, was its GPU. The Adreno cores buried somewhere in the chipset felt underpowered for anything heavier than Stardew Valley, and while DirectX and Vulcan compatibility was high, and driver updates refined the process, some games just didn't work.

With the release of the Snapdragon X2, a few things have changed. The main one is that the CPU package now comes with a lot more processing cores—18 on the Elite Extreme model we've got here—but the GPU is also a lot better. It's now putting out the sort of benchmark results you'd expect from Intel iGPUs, and in some tests can even produce better results than Panther Lake.

The ASUS ZenBook A16 we're looking at here is one of the first Snapdragon X2 laptops on the market. It boasts the Elite Extreme version of the chip, a name that makes you wonder where they're going to go next if they release an even better one. Elite Extreme Pro Max Gold Caviar Truffle Edition, perhaps. This one has 18 cores, and according to CPU-Z 12 of those are Performance cores running at 4.4 GHz, with the other six being Efficient cores running at 3.6 GHz. That's a lot of cores, and with ARM chips' ability to produce exceptional single-core performance (as shown by Apple's M5 last year) I had high hopes for this one.

And I was right. The X2 doesn't hit the same Geekbench 6 heights as the M5, but it's close, and beats the desktop Core Ultra 7 265KF and mighty Ryzen 9 9955HX in the single-core test. It beats those chips in the Geekbench multi-core test too, along with the Core Ultra 9 275HX from the Alienware 16 Area 51, and roughly the same as the 14-core M4 Pro in a 16in MacBook Pro.

The elephant in the room here is Panther Lake, which is going to be appearing in new laptops at the same time as the X2. And in Geekbench, the X2 and Core X9 388H both score over 3,000 points in the single-core test, with the X2 marginally in front, but in multi-core the gap widens, probably because the Intel chip only has four of the full-fat performance cores and 16 in total. It scores 17,924. The X2: 22,785. That's a difference of almost 24%.

Moving over to the Cinebench single-core CPU rendering test, which has been known to bring laptops to their knees and in which 100 is a respectable score, the X2 scores 151. A laptop running the X9 388H scores 130 in this test, and the multi-core rendering workout goes the same way, with the X2's 1391 beating the Panther Lake chip's 1173.

All this is assuming, of course, that the same tests applied to different CPU architectures can be directly compared. Geekbench has been designed for this, though as more of a tool for plotting broad relationships rather than a perfect one-to-one comparison between PCs.

All that single-core performance makes the ZenBook A16 very responsive, and if Microsoft's new Low Latency Profile makes its way to Windows on ARM it should make things even better. The fact that Asus has paired Qualcomm's X2 with 48 GB of fast RAM certainly helps too.

Things are looking quite good for the ARM chip until you remember it still uses an integrated Adreno GPU, and we have yet to see one that's paired with a discrete graphics chip. Still, it's not bad for an iGPU, supporting DX12 Ultimate and Vulkan 1.4, and putting out enough frames to play games in 1080p as long as you don't mind medium to low settings. Cyberpunk 2077 looks surprisingly good on medium, and while there are some shimmering textures in Blackmyth: Wukong, it doesn't struggle too badly as long as you remember to switch FSR upscaling on. Older titles such as Horizon: Zero Dawn look like what they are: PS4 games (albeit ones running at nearly 60 fps in this case, rather than the locked 30 fps of the console version). But it's great to see the Adreno is capable enough to play them as they were originally meant to be seen.

There are still some issues with the occasional game—F1 2024, for example (and this might be a different story with a more up to date release), putting up a notice that it's not compatible with ARM processors. This could also be because it's a fairly new processor at the time of writing, and a few software patches and driver updates will sort it right out.

On the outside, this is an unassuming laptop, though as a 16-incher it has a presence that you just don't find in smaller models. It's a kind of beigy grey, which isn't utterly unpleasant but certainly stands out from all the black and silver casings out there, and is made from Asus' Ceraluminum material that's some sort of unholy amalgam of aluminium and ceramics. It has a matte feel, and makes the A16 extremely light, but it's also stiff enough to handle by the corner without the trackpad becoming non-functional.

That trackpad is a big one, too, smoother than the surrounding casing but the same colour, and a better experience than the keyboard, which doesn't have much travel and is the worst aspect of this particular package. The screen, however, might be one of the best. It's a 2.8k 16:10 OLED, not a touch-sensitive one, with a 120Hz refresh rate. It's sharp, bright (500 nits), colourful and all the things a laptop screen should be. Battery life is exceptional—at least it is if all you're doing is looping a video, at which it kept going for almost 17 hours. Make it work a bit harder and that's going to drop, and there's a fan to keep it cool when it's under load.

So once again, Qualcomm has released a capable chip with a bit of a weak GPU, though it's nowhere near as bad as last time. Asus has then put it into a laptop with an excellent screen, though it is a little bit expensive when you could pick up something with a low-end GeForce chip for less, and have a superior gaming experience.

I can't help thinking though, that if someone could teach the X2 to talk to an RTX card, even as an eGPU, then we could have something truly fearsome on our hands.

This is the monitor I've been waiting for. As someone who has just entered another Counter-Strike arc that is sure to last an epoch, as my arcs tend to, I've been on the look-out for a 1080p monitor around the 360 Hz (ish) sweet spot that doesn't break the bank. Enter stage left: the KTC 25M1, at just $350. And I've already seen it on sale for under $300, which bodes well.

Of course, this price-to-refresh-rate ratio wouldn't matter if the monitor sucked, but I'm happy to report it doesn't. Far from it, I'm finding it a genuine pleasure to use for my late night Faceit matchmaking sprees. In what is proving to be both a blessing and a curse, gameplay is so smooth, clear, and snappy that I find it impossible to blame any blunders on the panel. Even if I could convince myself my opponents are using the $1,000, 600 Hz Zowie panel I recently tested, it wouldn't matter because the difference in practice is so slight that I highly doubt it would actually make my enemies any better.

The 24-inch, 1080p monitor, clad with detachable flappy side-wings (to prevent distraction), rear-side headphone stand, cable management feed, and the usual smattering of ports including a 3.5 mm headphone jack, is bred for one purpose: competitive shooters.

No matter how fast it is, you're probably not getting a TN panel for casual gaming or even competitive non-FPS gaming. Even though this panel's colour is damn good for TN, it's not as lush as an IPS and certainly not as vibrant as an OLED.

While it offers a solid 98% sRGB coverage, its DCI-P3 is only rated to 85%, meaning its HDR content is average at best, which is what I'd expect from a TN panel. That fits for my experience turning HDR on in Doom Eternal and eyeballing it: lighting seems a little more natural, but it's hardly night and day between that and sRGB mode.

The sRGB mode thankfully seems pretty well calibrated to my eyes. In addition to making casual gaming pleasant, this makes day-to-day use outside of gaming surprisingly comfortable, too—surprising, that is, for a TN panel.

HDR, though, is certainly not as good as what you get on the 320 Hz Alienware IPS panel I tested. Still, that was an IPS, not a TN. And more expensive esports TN panels like the 600 Hz Zowie one don't have it as an option at all.

For this monitor's real purpose, though, HDR is irrelevant. Sticking it in FPS1 mode but raising brightness a little, Counter-Strike displays flawlessly: enemies are clear and easy to spot, and movements are incredibly smooth.

I used an Nvidia Latency and Display Analysis Tool (LDAT) to test end-to-end latency—thus indirectly testing the response time—and it reported an average of 6.42 ms in-game over 150 muzzle flashes. That's well within a threshold where shooting feels and appears instantaneous to my eyes.

Latency performance

End-to-end latency

KTC 25M1

6.42

Alienware AW2525HM

6.09

Zowie XL2586X+

6.81

012345678910

Time (ms)

End-to-end latency Data

Product	Value
KTC 25M1	6.42
Alienware AW2525HM	6.09
Zowie XL2586X+	6.81

Tested using Nvidia's LDAT over 150 muzzle flashes in-game.

It's even quicker than the 600 Hz Zowie panel (6.81 ms), and only a little slower than the 320 Hz Alienware one with overdrive on (6.09 ms). These differences are all sub-millisecond, though, meaning they're imperceptible and for all intents and purposes we can say they perform about as well as each other when it comes to response times. I certainly didn't notice any difference in snappiness between the three when playing Counter-Strike.

For whatever reason, with this KTC monitor, I can't seem to find a mode that allows me to enable its 'Ultra Fast' overdrive, the fastest setting. But that's okay, because the middle 'Advanced' setting is snappy enough, and it's usually inadvisable to use the fastest overdrive mode in monitors anyway, as it risks inverse ghosting/halo effects. With the Advanced mode enabled, I haven't noticed any pixel overshoot.

Motion clarity is impeccable to my eyes, too, thanks to that 400 Hz refresh rate. As you can see from my slapdash moving photo of an also-moving UFO test, the high refresh rate really does make a difference. The top animation shown was running at 400 fps, the second at 200, the third at 100, the fourth at 50, and the bottom one at 25.

This difference is visible when eyeballing the moving alien targets, too: 400 Hz is just that much clearer than the rest. Admittedly, playing at 240 Hz in game will still be plenty clear, but this is just another example that goes to show what I've argued previously: around 360 Hz (or 400 Hz, in this case) is the sweet spot for competitive gaming.

To my memory, this monitor is also a bit clearer than I found the 320 Hz Alienware monitor to be when I tested it. I simply can't fault it for smoothness and snappiness.

The bonus here is you're getting the 360–400 Hz sweet spot for a damn good price. The 400 Hz Zowie equivalent costs $650, while this KTC monitor costs about $350 and can even be found for just shy of $300 when it's on sale. The 320 Hz Alienware monitor comes in cheaper than this, but honestly for the extra refresh rate and motion clarity boost, I'd say the KTC is worth it if you can afford it. If not, the Alienware will serve well.

With the KTC 25M1, you're hitting the refresh rate sweet spot just before diminishing returns kick in, which means you're getting some of the absolute best competitive/esports performance you can hope for before prices start to soar.

I feel KTC has struck the balance very well, keeping the price as low as possible for the performance on offer, and keeping things within a budget that doesn't only cater to pro players but also to your average competitive FPS gamer like yours truly. This would be the monitor I'd choose for competitive FPS gaming right now if my own money were on the line.

The Asus ProArt lineup is, in essence, the company's range of fancy PC products for creative professionals—and my goodness does it contain some nice-looking stuff. Which is why I jumped at the chance to test out one of the latest additions, the Asus ProArt Mouse MD301.

It's certainly a handsome squeaker, designed to provide a subtle, refined, ergonomic option for those who have some serious work to do. It's all business, this mouse, and there's something desirable about a design that eschews gamer-like frippery and instead aims for a more executive flair.

It's also customisable, in the sense that you can remove its main button switches for something more serene, or more clicky. There's also a horizontal scroll wheel, à la the Logitech MX Master range, which feels beautifully dampened. Really, most of the controls here are of the premium, well-thought-out sort, and that's very pleasing to see.

However, this $100/£105 professional rodent has the odd quirk that gives me pause. The side buttons are a touch flimsy. The thumb rest is slightly small and a little awkward. I've also had the odd connection issue, and the software feels like an afterthought.

There's a lot to like about the MD301. But as a long-term tester of the Logitech MX Master 4, I can't help but feel that if you're going down the professional ergo mouse route, it's the better buy of the two. The ProArt MD301 is a lighter, leaner attempt at an ergonomic mouse with many admirable qualities, but it drops the ball a little too often to give it my full seal of approval.

Asus ProArt Mouse MD301 features

The Asus ProArt Mouse MD301 is an ergonomically-shaped mouse that feels a lot smaller than your average ergo squeaker, coming in at 114.8 mm long, 75.2 mm wide, and 46.8 mm tall. Sit it next to a Logitech MX Master 4, and it looks positively tiny. It's also cheaper, coming in at $100 compared to the Logitech's $120 MSRP.

It's got six programmable buttons, including all the usual suspects. A left and right click, a middle scroll wheel click, two side buttons, and a hidden DPI button underneath a ridged plastic surface integrated into the chassis behind the wheel. It looks, in many ways, like a default mouse—albeit with some subtle ergo touches

Oh, and it's also got a horizontal scroll wheel mounted to the left-hand side. Logitech really changed the game with the old MX Master 3S design, didn't it?

Another interesting feature to note is the switches underneath the left and right buttons. By default, the ProArt MD301 comes with two (deep breath in) TTC E169-60BB0G0 (80M) units, but these push-fit socketed models can be pulled out and replaced with supported TTC, Omron, and Kaith units, which Asus has handily listed here.

The scroll wheel has a party trick, too, in the form of Asus' SmartShift tech. This allows for precise, clicky control when spun slowly, and a free-wheeling action when you spin it at an appropriate speed. It's a lot like Logitech's magnetic solution on paper, but feels quite different.

Spin the wheel at high speeds, and you can still feel it notch underneath your fingertip, albeit very quickly. Whereas Logitech's infinite scroll feature frees the wheel up with a glass-like, smooth action, the MD301 has a tendency to keep some tactile feedback in play, which makes it easy to stop at a defined point with precision.

Asus says the MD301 has up to 180-day battery life, and can gain eight hours' worth of charge in a single minute over its USB Type-C connection. Speaking of which, it also comes with a braided Type-C to Type-A cable, and a small USB Type-A dongle for wireless connection. Bluetooth connection is also supported.

Sensor-wise, this Asus squeaker makes do with an 8,000 DPI optical unit, which, in combination with its 1,000 Hz maximum polling rate, is unlikely to set gamers' hearts a-fluttering. Still, the MX Master 4 does just fine with a similar sensor (and a much lower polling rate), so it's not unsurprising for a work-first mouse. It's also a fair bit lighter than Logitech's latest ergo effort, tipping the scales at 99.7 g compared to the MX Master 4's 150 g bulk.

Asus ProArt Mouse MD301 ergonomics

The first thing I noticed about the MD301 is its diminutive proportions, especially when sat next to the MX Master 4. It's a lot narrower, with a much smaller thumb rest. However, it's got a wider top "hump", which feels good underneath my palm.

Switching between the two, it's clear that the Asus is designed to rotate your hand over to the right side more drastically than most, although at nowhere near as extreme an angle as something like the Keychron M5. It doesn't feel aggressively ergo, though. More like a slightly chonkier gaming mouse with some interesting quirks.

Clicking and spinning my way around the controls, I've been continually impressed by the effort that's gone into the switchgear and material choices. The left and right switches have a very positive click action, and well-weighted buttons.

The scrollwheel feels notchy yet free-flowing in all the right ways, with its clicky, fast-speed scroll being something of a highlight. The same goes for the horizontal wheel, which is heavily dampened and weighty underneath my thumb. Nice.

I'm also a big fan of the smooth outer coating, which has a silky quality that feels rather pleasant underneath my fingertips. It's slightly fingerprint-prone, but no worse than any other black gaming mouse I test these days.

The DPI switch is hidden underneath a ridged plastic insert that runs down the middle of the design, and provides a small moment of delight when you realise it's exactly where you expect it to be, but tastefully obscured. This is a classy mouse. A device for grown-ups. A refined rodent, if you will.

The only real ergonomic drawbacks come in two forms. One, the left-side thumbrest is quite small, and dips inwards. Using the MX Master 4 on the daily, my thumb has become accustomed to lolling on its own throne-like protrusion of silicone—whereas here it needs to tuck itself into a diminutive (and firm) plastic seat. My digit has become lazy, it seems, and has a tendency to want to fall off.

And two, the side buttons don't feel as high-quality as the rest of the unit. They're perfectly functional, but sitting as they do beneath a gorgeously weighted horizontal scrollwheel, the contrast is immediately obvious. They're simply "fine," while the rest of the controls are pretty excellent.

Overall, though, the ProArt MD301 is an impressive-feeling device, especially given its relatively low weight compared to its chassis size and ergo design. If anything, I could stand for it to be slightly heavier, in return for better side buttons and a bit of extra room for my thumb.

Asus ProArt Mouse MD301 performance

Given this is an ergo mouse with a relatively modest sensor, it would be unfair to expect Razer Viper V4 Pro-like performance or tracking ability from the MD301. Lo and behold, it's not the sharpest-performing rodent on the block.

Still, the sensor tracks well enough for work and light gaming usage, as shown by the MouseTester charts below.

I have, however, experienced the odd connection drop out. This has manifested as a very occasional half-second pause, and has happened several times during my week-long testing. I've got a lot of 2.4 GHz wireless devices plugged into my machine, and several more around my front room, so it's possible that the dongle here is slightly interference-prone—although I haven't been able to pin down the exact cause.

I have two other gaming mice plugged into this PC at all times (the MX Master 4 and the Razer Cobra Hyperspeed), and neither have had any connection issues. I tried unplugging both mice to see if it fixed the problem, and managed a few days of solid connection—before again experiencing a very quick pause.

I also had one connection pause while using the mouse with a gaming laptop I'm currently testing, so it doesn't seem to be specific to my main machine. The problem is so intermittent as to be barely noticeable, but there's something in this room the MD301's 2.4 GHz dongle doesn't appear to like.

Asus ProArt Mouse MD301 aesthetics

I love the ProArt MD301's aesthetic design. It's somehow utilitarian, slightly luxury, handsome, and hardly noticeable all at the same time. The ridged plastic vertical insert framing the scrollwheel is about as simple as design touches come, and yet adds a certain something that both feels good underneath your fingertips, and looks ever-so-slightly special on your desk.

Do I miss RGB lighting, or a touch of gamer flair? Not a bit of it. This mouse is all business, and that by itself makes it stand out.

Asus ProArt MD301 software

The ProArt MD301 connects to Asus Gear Link, a web-based app that allows you to adjust the DPI, polling rate, and button response settings within a browser. In order to tweak macros and use some of the features, though, you also need to install a companion app as an extension, which somehow feels like the worst of both worlds.

Pick either a web app, or installed app, Asus. Having one in order to properly use the other feels like an extra level of annoyance.

The web app itself has the odd quirk, too. I regularly split my 32-inch, 4K screen into two side-by-side browser windows, but the Asus app complains that it doesn't have enough room in one panel, suggesting I "switch to a larger screen." What do you want, a 49-inch ultrawide?

There's an option to use it anyway, though that's a new one on me. It seems to scale the interface just fine if you override it, but it's a bit indicative of the software experience overall. Unfinished is the feeling I come away with.

Still, early annoyances aside, the web interface works well enough. It's a bit barebones, though, and given the premium-feel of the mouse itself, it strikes as a clunky afterthought tacked on to what is otherwise a well-put-together package.

Asus ProArt MD301 value

At its $100 MSRP, I'd say the Asus ProArt Mouse MD301 is well-priced for its bracket. It's a chunk cheaper than the Logitech MX Master 4, and looks and feels like a premium mouse for more reasonable money.

That being said, I can't help but feel you're getting a lot more for your cash with Logitech's offering. It's got a more comfortable thumb rest, near-silent switches as standard, and a set of controls that can match (and occasionally beat) the Asus in terms of feel. And, it must be said, the Logitech software experience is a lot more refined. Plus, I've not experienced any connection pauses with the Master in many months of everyday usage.

Which puts the ProArt squeaker in an odd spot. It's a lot lighter than the Master 4, which makes it more suitable as a gaming mouse. But it's still got an ergo-spec sensor, so I doubt many will be buying it for its gaming performance chops.

And while it's a good-looking object that would suit the *loudly coughs* executive-ergo-mouse-using-professional, I can't help but feel there are enough hang-ups here to cause them to sniff into their yuzu-infused lattes.

It's a good effort from Asus, and a mouse with many likeable qualities. However, the quirks slightly outweigh the advantages, and as a result, I struggle to recommend it over the competition.

The LiberNovo Omni Pro is what happens when a very good first‑gen idea goes in for a polish rather than a full redesign. If you've read my original LiberNovo Omni review, you know the bar here is high. That chair is still the comfiest thing I've ever sat in for ridiculous work‑and‑gaming days. The Pro turns that same formula up a notch in a few key areas—materials, finish, and a couple of tech tricks—but it doesn't fundamentally change how it feels to sit in.

Day to day, comfort is still excellent. The cushions are plush but supportive, and I can go from writing to gaming without my back complaining. LiberNovo re‑engineered the seat with multi‑density foam, but if I'm honest, I can't say I feel a big difference. The original already nailed it, so the Pro feels more like a refinement than a really tangible upgrade here.

The bigger change is how well that comfort can now fit your body. You can choose a 45 cm or 48 cm seat depth at checkout, which is great if you've ever had a chair either cut into the back of your knees or stop awkwardly short of your thighs. Average‑sized folks will be fine on 45 cm; if you've got longer legs, 48 cm gives you more support under there. And if you're properly tall or need something truly built for bigger frames, LiberNovo's Maxis line is really the one you should be looking at first.

The big change you notice while actually using the chair is the new five‑position recline system: 105°, 115°, 125°, 135°, and 160°. Each angle is pitched as a "mode." Deep Focus at 105° is your upright, camera‑on posture. Nudge it back to 115°—Balance mode—and your spine and shoulders relax a bit without dropping you into lazy‑boy territory.

At 125° (Solo‑Work), you're in that natural lean‑back zone that suits long editing or deep‑work sessions. 135° (Soft Recline) is where I ended up with a controller in hand, still close enough to the desk to reach controls without overextending. Then 160° is Spine Flow, the full lay‑flat angle for power naps and the motorized OmniStretch spinal stretch.

On paper it's a smart, lifestyle‑based ladder of positions. In practice, the tuning isn't perfect. The jump from 105° to 115° feels a little abrupt and less smooth in motion than the original Omni's recline. Once you're actually settled into Solo‑Work or Soft Recline it feels fantastic, but it does give off a bit of "we added more stops for the spec sheet" energy rather than solving a problem I actually had.

Where the Pro does clearly earn its name is build quality and materials. Overall construction feels a notch tighter than the original: cleaner stitching, subtle design flourishes on the wheelbase, and noticeably less creaking from the backrest when you shift around. It feels less like a Kickstarter experiment and more like a polished production chair.

The headline upgrade is that Gabriel Atlantic Danish fabric. Terrible branding, fantastic textile. It looks and feels genuinely premium—more high‑end task chair than shouty gaming throne—and it has the numbers to back it up: abrasion resistance rated for roughly a decade of heavy use, OEKO‑TEX certified, breathable, flame‑retardant, and machine‑washable without shrinking. In the real world, the question is how it handles crumbs, spills and kid‑induced chaos, and that's something only months and years of abuse will reveal. Right now, it feels like the right kind of overkill and looks stunning at my white Secretlab Magnus desk.

Ergonomically, most things are evolutionary. The 4D armrests are a bit less prone to spinning or drifting when you bump them, though they're still not perfect. The headrest has more adjustability and is lovely when you're reclined, but in a true upright focus posture it still doesn't come out far enough for my liking. It feels tuned first for lounging, second for locked‑in work.

The motorized lumbar system has been revised internally and should be more durable. The feel, though, is very similar to the OG Omni: you get depth adjustment but not height. If the default strike point doesn't line up with your lower back, it can still feel more like a mid‑back nudge than that sweet, targeted lumbar hug. In the reclined modes it's more forgiving; in upright, I still wish they'd added vertical adjustment.

Then there's the new party trick: Active Airflow Seat Ventilation. Under the seat is a centrifugal fan pushing air up through perforated foam and mesh, with two levels—Low and High—rated for up to 36 and 9 hours respectively off the built‑in 3000 mAh battery. There's a seat sensor so it automatically pauses when you get up and resumes when you sit back down, plus a memory function so it remembers your last setting.

Here's the thing: it's winter where I am, so I haven't been desperate for a cooled backside. When I did use it, the sensation is more of a gentle coolness than "there's a fan under my bum." I can see this being a genuine quality‑of‑life upgrade in summer or if you work somewhere hot, but right now it feels more like a clever extra than the killer feature that justifies the "Pro" price tag.

The rest of the electronics story is thankfully boring in the good way. Battery life remains excellent; even with mixed use—occasional lumbar tweaks, the odd Omni stretch cycle, some airflow thrown in—I never got close to killing it. Like my Omni, I don't expect to be charging this one very often which is a great convenience.

What hasn't changed is the warranty, and that's the part that continues to bug me. You get five years on the frame and only two on the electronics, even though all the Pro‑specific magic lives in those electronics. LiberNovo promises better long‑term reliability this time around, but until these have been out in the wild for a few years, that's just a promise. And the optional footrest? Still the weak link. I didn't rate it on the original and it's effectively unchanged here: okay for lounging, not something I'd build my daily ergonomics around.

So where does that leave the Omni Pro? It's clearly the flagship of the range. You're paying extra for nicer fabric, cleaner finish, a more articulated recline and the cooling seat. Comfort, however, hasn't really moved on. This is still one of the most comfortable chairs I've used, but it doesn't feel noticeably comfier than the original Omni. The gains are mostly in polish and feel rather than in raw, "my back is changed forever" ergonomics.

Does it really earn the "Pro" badge? Sort of. It's the best, most premium expression of the Omni idea, not a ground‑up reimagining. If you want the nicest version and you're happy to pay for upgraded materials and the tech flourishes, the Omni Pro will absolutely treat you well. If you're more budget‑conscious, or simply less excited about motors and fans in the thing you sit on all day, the Omni SE suddenly looks very smart: you lose the electric lumbar and airflow, keep the core comfort and ergonomics, and save around $400. For a lot of people, that's going to be the real sweet spot.

Honestly, I want to like the Geekom A9 Max, I really do. From a pure design perspective, the hardware, the engineering behind it, the connectivity bundled into this thing, it's on another level. And it's just a bloody pretty thing at that, at least as far as Mini PCs go. You'd think it'd be a shoo-in for our best SFF PC list, without question. But there's a catch, because of course there is.

I may as well rip the band-aid off right now. It's the memory. I can already see the PR team squirming as they read that sentence, and it genuinely does pain me to say it, I promise. Geekom has decided to ship the A9 Max AI with a single stick of 32 GB DDR5 @ 5600 MT/s. And that is a decision that has effectively starved AMD's Radeon 890M of any modicum of the performance it can actually generate. I know this because I have extensively tested this hardware most recently, in Acemagic's Mini PC Retro X5, which, like-for-like, practically has the same spec, bar a smaller SSD, and a slightly older CPU.

The performance impact that this one tiny decision has had on the overall architecture is nuts, and combined with the wildly high price tag being what it is, it's by far the thing that's decimated that review score. If Geekom turned around tomorrow and popped 2x 16 GB sticks in this config instead of this single-channel nonsense, I can happily tell you right now, the score would bounce up by a good 15%, because in almost every other way, the A9 Max is excellent.

And, breathe. I'll get back to that in a minute. Still, for £1,600 in the UK or $1,800 US, it's generally not a bad offering (memory aside). You get that 32 GB RAM capacity, a 2TB Kingston PCIe 4.0 SSD (it's not the best), AMD's Ryzen AI 9 HX 470 CPU, A Radeon 890M iGPU, and a hell of a lot of connectivity.

The chip itself is relatively new on the scene, only really landing with us in January of this year. Though it's effectively a refreshed Strix Point (HX 370 series) chip, still built out of the same silicon architecture as its predecessor. You get those four Zen 5 cores, and eight Zen 5c cores for a fat 12/24 multi-threaded config, and it's all still built on the back of TSMC's 4 nm process as well. What's different? Well… clock speeds have been bumped ever so slightly up (with an extra 100 MHz on boost), oh, and the "biggest" update comes in the form of the NPU, which has seen its performance flick up to 86 TOPS versus its predecessor's 80 (huzzah, our AI datacenter worries are finally solved).

Connectivity is wild, too. There's so much here if you're looking for a lightweight workstation that's not brimming with a dedicated GPU and full-fat processor. There's no less than five USB 3.2 Gen2 Type-A, two USB 4.0 Type-Cs, twin 2.5 Gbe, along with support for twin HDMI 2.1 ports, as well as the other bells and whistles you'd expect. No Thunderbolt, of course, but you do get WiFi 7 and Bluetooth 5.4.

Alright, I'm going to give you some numbers now, and they're quite clearly bad. But, it's important that we all collectively have context on this before I act like a wailing harpie and impale this memory situation into the dirt with my shiny shiny words (honestly, it's a miracle I get paid for this).

So, we test our SFF PCs (with iGPUs) at 1080p on the medium graphical preset. For this little wee dissection, I'm going to be using that Acemagic X5 I mentioned earlier for comparison. It has the same Radeon GPU, the same memory speed, memory capacity, a better SSD (albeit smaller), and practically the same CPU for what it's worth (technically, the A9 Max has a better chip here). On average, across all our test titles, with and without upscaling, across five separate games, separate engines, separate developers, the A9 Max landed just 23.1 fps. The Retro X5, with its dual-channel memory, 47.6 fps.

Previous

1 / 3

Next

1080p gaming performance

Black Myth Wukong (1080p Medium)Cyberpunk 2077 (1080p Medium)F1 24 (1080p Medium)Horizon Zero Dawn (1080p Original)Metro Exodus Enhanced (1080p High)

Avg FPS

1% Low FPS

Geekom A9 Max

9

7

Acemagic Retro X5

20

16

Black Myth Wukong (1080p Medium) Data

Product	Value
Geekom A9 Max	9 Avg FPS, 7 1% Low FPS
Acemagic Retro X5	20 Avg FPS, 16 1% Low FPS

Avg FPS

1% Low FPS

Geekom A9 Max

22

15

Acemagic Retro X5

38

22

Cyberpunk 2077 (1080p Medium) Data

Product	Value
Geekom A9 Max	22 Avg FPS, 15 1% Low FPS
Acemagic Retro X5	38 Avg FPS, 22 1% Low FPS

Avg FPS

1% Low FPS

Geekom A9 Max

30

24

Acemagic Retro X5

73

59

F1 24 (1080p Medium) Data

Product	Value
Geekom A9 Max	30 Avg FPS, 24 1% Low FPS
Acemagic Retro X5	73 Avg FPS, 59 1% Low FPS

Avg FPS

1% Low FPS

Geekom A9 Max

23

13

Acemagic Retro X5

43

32

Horizon Zero Dawn (1080p Original) Data

Product	Value
Geekom A9 Max	23 Avg FPS, 13 1% Low FPS
Acemagic Retro X5	43 Avg FPS, 32 1% Low FPS

Avg FPS

1% Low FPS

Geekom A9 Max

14

5

Acemagic Retro X5

23

8

Metro Exodus Enhanced (1080p High) Data

Product	Value
Geekom A9 Max	14 Avg FPS, 5 1% Low FPS
Acemagic Retro X5	23 Avg FPS, 8 1% Low FPS

You can clearly see how a memory bandwidth starved iGPU is hobbled by having single channel memory here.

1080p upscaled gaming performance

Black Myth Wukong | Upscaling (1080p Medium)Cyberpunk 2077 | Upscaling (1080p Medium)F1 24 | Upscaling (1080p Medium)Horizon Zero Dawn | Upscaling (1080p Original)

Avg FPS

1% Low FPS

Geekom A9 Max

14

11

Acemagic Retro X5

30

22

Black Myth Wukong | Upscaling (1080p Medium) Data

Product	Value
Geekom A9 Max	14 Avg FPS, 11 1% Low FPS
Acemagic Retro X5	30 Avg FPS, 22 1% Low FPS

Avg FPS

1% Low FPS

Geekom A9 Max

30

20

Acemagic Retro X5

52

27

Cyberpunk 2077 | Upscaling (1080p Medium) Data

Product	Value
Geekom A9 Max	30 Avg FPS, 20 1% Low FPS
Acemagic Retro X5	52 Avg FPS, 27 1% Low FPS

Avg FPS

1% Low FPS

Geekom A9 Max

30

26

Acemagic Retro X5

86

75

F1 24 | Upscaling (1080p Medium) Data

Product	Value
Geekom A9 Max	30 Avg FPS, 26 1% Low FPS
Acemagic Retro X5	86 Avg FPS, 75 1% Low FPS

Avg FPS

1% Low FPS

Geekom A9 Max

36

21

Acemagic Retro X5

63

44

Horizon Zero Dawn | Upscaling (1080p Original) Data

Product	Value
Geekom A9 Max	36 Avg FPS, 21 1% Low FPS
Acemagic Retro X5	63 Avg FPS, 44 1% Low FPS

The performance gap between the system with dual-channel memory and the one with single-channel memory doesn't get any better with upscaling enabled.

Synthetic gaming performance

3DMark Time Spy3DMark StorageFFXIV Shadowbringers

Index score

GPU index score

CPU index score

Geekom A9 Max

2301

2035

8948

Acemagic Retro X5

3532

3153

11109

3DMark Time Spy Data

Product	Value
Geekom A9 Max	2301 Index score, 2035 GPU index score, 8948 CPU index score
Acemagic Retro X5	3532 Index score, 3153 GPU index score, 11109 CPU index score

Index score

Avg bandwidth (MB/s)

Access time (ųs)

Geekom A9 Max

1306

226.22

139

Acemagic Retro X5

2460

415.65

72

3DMark Storage Data

Product	Value
Geekom A9 Max	1306 Index score, 226.22 Avg bandwidth (MB/s), 139 Access time (ųs)
Acemagic Retro X5	2460 Index score, 415.65 Avg bandwidth (MB/s), 72 Access time (ųs)

Geekom A9 Max

38.96

Acemagic Retro X5

20.45

FFXIV Shadowbringers Data

Product	Value
Geekom A9 Max	38.96
Acemagic Retro X5	20.45

Contrary to what Geekom says, the synthetic performance difference is actually less than real-world in-game benchmarks.

It's not a limitation that sits solely with gaming, either. In 7-Zip, compression and decompression were markedly slower, multi-core performance in CineBench was down too, and the only times it did beat out the HX 370 was in single-threaded tasks, where memory bandwidth is less of a thing. Not exactly a riveting score sheet then. Even 3DMark's Storage benchmark came out significantly slower, which I can't believe even for a second is entirely down to the drive being just that much sloppier than the EXT X200E found in the X5 is.

Previous

1 / 2

Next

Synthetic CPU performance

Cinebench R247-zip

Single-core index score

Multi-core index score

Geekom A9 Max

119

1101

Acemagic Retro X5

107

1169

Cinebench R24 Data

Product	Value
Geekom A9 Max	119 Single-core index score, 1101 Multi-core index score
Acemagic Retro X5	107 Single-core index score, 1169 Multi-core index score

Compression (GIPS)

Decompression (GIPS)

Geekom A9 Max

98.88

130.12

Acemagic Retro X5

110.45

131.4

7-zip Data

Product	Value
Geekom A9 Max	98.88 Compression (GIPS), 130.12 Decompression (GIPS)
Acemagic Retro X5	110.45 Compression (GIPS), 131.4 Decompression (GIPS)

This is likely where Geekom's claims about limited performance difference ring true, as you can see both systems performing well.

Cooling performance

Thermal performance

Max CPU temp (°C)

Max CPU package power (W)

Geekom A9 Max

86

64.98

Acemagic Retro X5

85

69.98

Thermal performance Data

Product	Value
Geekom A9 Max	86 Max CPU temp (°C), 64.98 Max CPU package power (W)
Acemagic Retro X5	85 Max CPU temp (°C), 69.98 Max CPU package power (W)

This is likely where Geekom's claims about limited performance difference ring true, as you can see both systems performing well.

Just to be doubly sure, I re-ran the whole benchmark suite, and triple-checked GPU frequencies, temps and power draw throughout that process too, noting down average frequencies as well (not typically something we report on for these machines). The X5 maxed out at 2,764 MHz and 2,100 MHz on average, at 60 degrees and with 50 W on max draw. The A9, 3,100 MHz at max, 2,300 MHz on average, with a top temp of 63 degrees and 55 W of power usage. By all accounts, then, the GPU itself was running at full tilt, but was just starved of memory bandwidth.

Geekom knows about this, too. In fact, I was pre-briefed that this was a decision it had made in advance. The justification being that it provides end-users with the option to expand their memory solution later down the line with an additional stick of 32 GB. Which, yeah, okay, I don't necessarily disagree with.

Geekom also notes that, actually, the performance drops are only in synthetic benchmarks, and don't affect real-world performance. Which I do disagree with.

You can quite clearly see that this does massively impact real-world results. In fact, more so than in the synthetic tests, by my calculations. What's particularly frustrating about all this is that nowhere on the product pages (at least at time of writing) does it suggest that the A9 Max ships with a single stick. It's all "dual-channel compatible", "up to 128 GB", or "32 GB RAM", "High-speed memory".

It's a difficult thing to just accept, because the truth of it is the A9 Max is an outstanding piece of design. It's cool, quiet, and it's clear that the CPU in particular is seriously potent, particularly in localised AI applications if that's your jam. The sad reality is, we've already seen what this hardware can do; it just came in a beige box at half the price.

Assassin's Creed Black Flag Resynced might go down as one of the more puzzling remakes ever conceived. Not only is it bizarre that Ubisoft decided to kick off a likely streak of remakes with the sixth game of the series, but it also chose a game so recently and universally enjoyed that the only way to "remake" it is to simply put out the same game again.

A cleaned-up remaster might have been better than what Ubisoft has done here. Resynced is tirelessly faithful to the qualities that set Black Flag apart in this series—sailing the Jackdaw is as great as ever, now distinguished by a newer Anvil engine that enables dynamic weather, time-of-day, and a truly seamless open world—but fundamental changes to combat, stealth, and encounter design reveal that this is essentially Black Flag stuffed inside Assassin's Creed Shadows.

Resynced is good. It's slick, ridiculously beautiful, and improves on the original Black Flag in several ways. I'm impressed by Ubisoft's Singapore's obvious love for the source material—every story beat and cutscene has been recreated shot-for-shot with original performances intact. If you've never played the second-best Assassin's Creed game ever made, Resynced will get the job done.

For a while, I even believed it would outshine the original Black Flag—I was wooed by Resynced's breathtaking Caribbean vistas that stretched further than my PS3 could be asked, taken by its swashbuckling ship-to-ship showdowns, and impressed by the return of "advanced parkour" mechanics like back and side ejects. Ubisoft spent a generation transforming Assassin's Creed from a stealth-action power fantasy into a loot-driven RPG with spammy combat, and only now is it trying to make amends. Resynced feels like Ubisoft teaching itself how to make a proper Assassin's Creed again, and graded on that curve, it's quite good.

But then, after rolling credits, I reinstalled that original game. Even though I was retreading missions that I'd literally just played newer versions of, I had so much more fun in Black Flag 2013 that I just kept playing… and playing, and playing, until I'd completed both games front-to-back in the same week.

This turned out to be a useful exercise (I told my fiancé, puzzled by a second round of sea shanties emanating from my Steam Deck), because it took a return to old Assassin's Creed to properly appreciate mechanics, design choices, and an entire storyline that, according to internet consensus and Ubisoft, weren't worth preserving. Now having braved Black Flag's dreaded tailing missions, indulged in Ezio-era chain killing, and hacked the computers of my Abstergo coworkers, I reckon this 13-year-old game has aged quite gracefully.

I'm glad Resynced is here, but honestly? The original is better.

Jumper

Take parkour and climbing. Ubisoft is making a big deal about Resynced's parkour being a step up from Shadows, Mirage, and Valhalla, and while that's true, that's setting the bar on the floor. Compared to the original, Edward's updated movement is stilted and floaty. Classic maneuvers are there, but they don't flow very well, resulting in freerunning that looks like bespoke animations playing on a timeline rather than a fluid sequence. Whatever black magic Ubi used to hide these animation seams in the 2010s was lost in the remake process.

Perhaps Ubi is satisfied with less convincing parkour as long as it's faster. Resynced's locomotion is designed as if Black Flag's meticulous animations were a nuisance. Remake Edward wallruns straight up single-story buildings like he's Spider-Man, then proceeds to expedite larger climbs through silly-looking vertical leaps that have more in common with Sly Cooper than Ezio Auditore. When I should be punished for chancing an unwise jump, Edward suddenly gains springs in his boots to clear the gap.

For as much guff as the old AC games get for finicky climbing, Black Flag represented six games of refinements—steering Edward over rooftops, through branches, and across highwires in 2026 is still expressive, and less prone to accidents than the first few games. Climbs are smooth, but weighty and methodical. It takes some time and a little brainpower to crest a church tower or fort. In my unscientific test using the same building in 2013 vs 2026, remake Edward flew up this church twice as fast:

I find the part where original Edward carefully climbs around a window overhang, followed by remake Edward phasing straight through it, to be illustrative.

Resynced is functional and quick, but it's not cool or (dare I say) immersive. And man, that used to be a core appeal of this series. Neither system is complex, but the old climbing at least feels intentional—like we're meant to experience the effort and time it takes to scale a structure that was never meant to be climbed.

That pace has a grounding effect that extends beyond walls. Remake Edward no longer runs with any simulated momentum, meaning he can instantly change direction or flip a u-turn like he's a Dark Souls protagonist. It's a very artificial, videogame-y feel that I ended up resenting the more I played the weightier original.

Blade dance

Even when Resynced's systems are objectively better, Ubi has failed to respect the rule of cool. Black Flag's reimagined combat carries over Shadows' dodge/parry and light/heavy attacks, with the addition of Edward's signature chain kills. It's more engaging than the original's simplistic "counter and instakill everybody" approach, but confusingly, the remake actually reduces Edward's combat options by a lot.

Hidden blades are no longer a primary weapon, now relegated to finisher animations. You can't steal enemy muskets, sabers or heavy axes—temporary pickups that carried their own advantages, drawbacks, and cool finisher animations in the original.

No, remake Edward is strictly a cutlass man, except now he can kick enemies. I don't think the tradeoff was worth it, because Black Flag was never lacking challenging combat. I remember liking Edward as a terrifying pirate terminator soloing the crews of rival brigs, and that impression was backed up by my replay. It's not cerebral stuff, but the weapon pickups gave me more ways to play with my food, and fluid chainkills are gratifying for the same reasons the Batman Arkham games still kick ass.

With respect to some new, freshly gruesome Edward finishers, the remake combat is too mushy and spammy to accomplish the same fantasy. Old AC is about messing dudes up and looking awesome doing it, but new AC is about cleaving away at a health bar.

This is especially true in Resynced's occasional forced boss fights, which replace moments where you're meant to stealthily assassinate targets. You know, like assassins do.

The chase

At least Resynced wins in the sneaking department. The arrival of Shadows' crouch button instantly elevates both the complexity and fantasy of stalking targets from rooftops. The majority of Black Flag's missions involve infiltration and neck-stabbing, so this is a significant upgrade. I felt empowered to explore compounds and avoid guards more often than I killed them, which I attribute to a gracious window between guards seeing Edward and actually pursuing him. Stealth within compounds was the highlight of Shadows, so this is no surprise, but I'm disappointed that Ubi has also repeated that game's mistakes with the cities surrounding those compounds.

The remakes of Kingston, Havana, and Nassau are faithful and, in some cases, even expanded, but Ubi has rendered them less dynamic by nerfing the British and Spanish troops that patrol them. Rooftop guards are completely absent in neutral areas, which kills all the tension of freerunning from A to B. If you do manage to piss off guards, they won't even try to chase you outside of their immediate station. Caught trying to steal booty in a restricted zone? Just run outside and they'll forget you exist.

Ubi has lost the art of the chase, and it sucks. Chases were once a fun, unique feature of Assassin's Creed that forced you to take paths that guards could not easily follow, break line-of-sight to avoid gunshots, and keep your cool long enough to lose a tail in a haystack, bench, or closet. Ubi had already started to deemphasize notoriety systems by the time Black Flag came around—guards will quickly forget about Edward once he loses them, and you don't have to tear down posters of your face—but chases themselves persisted. Resynced lovingly recreated all of those benches, wells, and closets, then gave me exactly zero reasons to use them.

It's one of many ways that Resynced subtly changes the original Black Flag in the name of "quality of life," but ends up blander for it. The series' infamous tailing missions are now much easier. Thanks to no rooftop guards, you can just follow from above with zero worries, and if you do get spotted, just skip to the end of the mission by killing your target and looting the relevant info instead. I know these missions were unpopular with the impatient, but the occasional round of enforced stealth gave AC's repetitive infiltrations some variety while utilizing city space that was otherwise just there.

They also reinforced that, occasionally, stealth is the entire point. These games aren't as fun when every problem can be solved by a hidden blade.

Ship combat has been grazed by the easy beam, too. The Jackdaw is significantly more maneuverable in the remake, a small but consequential player buff that makes follow-up cannon volleys easier while in motion. I also noticed that enemy ships were generally less aggressive on normal difficulty—when I returned to 2013 Black Flag after 25 hours with the remake, I got my ass handed to me by a level one pirate hunter.

The Abstergo of it all

And, oh yeah, Resynced also surgically removes a huge chunk of Black Flag's story. The entire conceit of the original game is that you're an Abstergo Entertainment employee rummaging through Desmond's DNA so that a fictional Ubisoft can turn Edward's memories into a videogame while the Templars covertly seek the precursor technology at the heart of the story. Those sequences are just totally gone, replaced by an Animus interface only used for battle pass unlocks. What a downgrade.

I'm surprised that the removal of all Abstergo business was universally cheered upon announcement. The present day stuff was never my favorite part of these games, but I've always considered Black Flag's a net positive. The interludes were short, semi-optional, and rewarded fans with a mountain of worldbuilding through emails, voice logs, and hidden videos. My recent playthrough reminded me that the ending of these episodes is pretty terrible, but so reveals a lost opportunity. Ubisoft could have directed resources at making this integral part of the original game better, but they just deleted it and replaced it with nothing at all.

It's revisionist history that nobody liked this part of the game—I did, for one, and plenty of reviewers did too. Are we meant to take this as Ubi retconning this series as nothing more than historical action games? Assassin's Creed is a sci-fi world, damn it!

Back-to-back

50 hours across two back-to-back Black Flag playthroughs left me with a lot to chew on. It's a testament to Black Flag's triumphant sailing that it was fun enough to do twice with minimal changes. Time has not been as kind to Edward Kenway, who I'm now convinced is a dull character with an unearned hero arc I was not smart enough to see through at 17 years old.

But the long gap between playthroughs was also long enough to miss a version of these games that was getting a bit tiresome by 2013. There's a comfort to the original's plodding pace, myriad collectibles, and simplistic swordplay that actually stands out more in 2026 than Resynced's "break yet another posture meter" combat and entirely optional stealth.

Resynced's advantages are plain to see. Its world is one of the most gorgeous and realistic depictions of nature ever put to videogame, and its simplifications will likely prove broadly appealing. This is also the better PC port, offering loads of graphical options and features with relatively chill system requirements. It's even Steam Deck verified at launch, though if it's roughly as performant as Shadows, expect a jittery 30 fps target. The original, on the other hand, is finicky in ways that many early 2010s console games were on PC. There's no borderless fullscreen, it doesn't like being alt-tabbed, and will hang for 20 seconds on startup trying to load ancient Uplay stuff unless you cancel it each time. It's also locked at 60 fps by default, but there's a mod for that.

Photoreal graphics are nice (especially on my RTX 5090/9800X3D machine), but man, have you seen the old Black Flag lately? The faces are rough, but that world is still a looker. It's a bit more stylish and grimy than Resynced's polished look, and I prefer it for a pirate adventure.

I just don't think Resynced's marquee features are all that important to me. I kept going through this exercise with Resynced where a new feature would be introduced and I'd assume it was new to the remake—"Oh wow, a home I can customize and improve over time! Neat, they added more ship ammo types! I don't remember these Mayan statue puzzles. Naval contracts? That's definitely new"—only to realize all of them existed in the original, untouched by the remake process.

Resynced is decent, but it takes away more than it adds, as if Ubisoft wasn't confident in the hit game it made 13 years ago. It's a conflicted remake that's worth a look, but is ultimately inessential.

The humble mouse has been through some changes in its time. When it’s not adding extra buttons all over the place it’s twisting your wrist into unlikely positions or looking like whatever the Mad Catz RAT was. A general theme over the years, however, is that they’re getting bigger, and perhaps there's a study to be done into whether PC gamers’ hands are getting larger too, the better to cradle and envelop them.

Logitech has taken the opposite approach with the Mobi—it’s truly tiny. This is a travel mouse, a mouse for the truly short of space, a sliver of silicone-wrapped plastic that folds in half when not in use, looking rather like the Surface Arc mouse Microsoft put out in 2017, but a little bit cheaper. Where that device folded flat for transport, this one goes the other way.

Hand the Mobi to someone and the conversation goes like this: “It’s a phone?” No. “It’s a phone stand?” No. “What is it then? I can’t believe I don’t know.” Click click click on the buttons. “Oh, it’s a mouse!”

While it may resemble a Star Trek communicator or a tiny flip phone to go with your geek pie hairstyle (“that’s well weapon”), the Mobi is in fact just a mouse. It has two main buttons, two further programmable buttons, a little touch-sensitive panel to act as a scroll wheel, a PAW3222 4,000 DPI optical sensor, and a USB-C port for charging. It connects to your PC (and two other devices) via Bluetooth with fast pairing, and will play nicely with a Bolt receiver if you ask it to, though one isn’t included in the package unless you buy a special ‘business’ version.

The folding nature of the mouse is actually a stroke of genius. When collapsed, it will slip neatly into your skin-tight futuristic space-catsuit without causing too much of a bulge. And when stretched out it fits in the hand, something many compact mice that just take a normal-sized one and shrink it can’t claim. Your third finger and thumb will go underneath the 120-degree (ish) central hump if you let them, and there's certainly space for them to do this, but most of the time there's just enough height in the body of the mouse for your thumb to rest on.

Unless you have insanely long fingers you won’t run out of space on top, and the third and fourth buttons are cleverly integrated under the scrollpad, with a press at the top or bottom activating them. These click much more loudly than the nicely damped main buttons, and you can use the Logi Options+ software for a degree of customisation about what they do, including a neat double-tap auto-zoom.

And while it may look a little flimsy, Logitech has some remarkable claims about how it’s not. It can, apparently, withstand 50,000 folding actions before it starts to wear out, which Logi thinks should take 15 years (or about 5,479 days, for an average of nine folds a day). Obviously that doesn’t take into account people constantly folding and unfolding it because of the novelty of the hinge mechanism, but in 15 years we’ll be controlling our PCs with brain implants anyway. Instead of a simple switch in the hinge to detect when it’s closed and switch it off, Logitech says there's an on-device AI that protects against unintentional clicks while you’re folding it. If that sounds absurd, it’s because it is, but that’s where we are in the space year 2026, requiring an AI to tell us whether we’re wearing pants or not.

Better is the battery life of 30 days from a full charge, and the ability to use it for 22 hours after a charge of just one minute. You won’t be able to use it while it’s charging thanks to the position of the USB-C port, which isn’t quite up to Apple Magic Mouse levels of egregiousness, but with a cable connected it feels enough like you’re about to snap the plug of in the socket to make you reconsider after a few seconds. The battery cover can be pried up with a screwdriver, and it looks like the cell could be replaced, though Logitech hasn’t mentioned anything about this.

Of course it’s possible to play games with it. You can game with a trackpad or the tiny red nub on a Thinkpad, and I’ve played Age of Empires II with much worse things than the Mobi. It doesn’t require any kind of special surface to glide over, seems quite happy on a trouser leg, and the pointer is reliable and accurate. The important thing here is how much better it is to use than a laptop’s trackpad, especially if you’ve not put in the hours required to get good at using one for gaming.

Sadly, that’s almost all the praise it’s possible to muster about the Mobi. Its four buttons are fine, but scrolling is a mess. Not only is the touchpad the same texture as its surroundings, so it’s quite possible to be lovingly rubbing your fingertip against one of the buttons instead of where the action is in the middle, but even if you do get it in the right place, the web page or camera zoom you’re controlling moves jerkily, as if it’s losing track of your fingertip.

A hand that’s used to a solid-bodied mouse can rebel against being asked to cradle something so slender, and you end up moving from what should be a straightforward palm caress to a crab-like grip on the edges. It’s possible to pick the back edge up, fold your fingers underneath, and keep only the sensor in contact with the tabletop, and this might be the best way to use it despite every promotional image showing it with front and back firmly on the ground.

So, yeah. The Logi Mobi Fold is a nice piece of engineering but inevitably can’t hold a candle to a proper gaming mouse. Perhaps that’s not the point, as we don’t expect products designed for extreme portability to perform precisely like their full-size counterparts. At 79 g it’s heavier than the Logitech G Pro X Superlight and the Corsair Sabre V2 Pro, so if you’ve got it tucked in your bag and fancy a bit of Total War it’s better than a trackpad, but a modern full-sized mouse is hardly a large piece of equipment to transport, so use that instead.

You can't have too many screens. That's pretty much been the guiding moral in my life. Enter the new Asus ROG Strix XG129C and a minor existential crisis. This is a screen, but I have to ask—I can't believe I'm saying this—is there any real point to it?

For starters, this "gaming monitor" (Asus' words, not mine) measures just 12.3 inches. So, it's pretty puny. Actually, it's worse than that. It sports an ultra-wide 24:9 aspect ratio and thus isn't even as big as those 12 inches might imply. In terms of vertical viewable screen height, we're talking sub-4.5-inch.

But here's the thing. While I came in decidedly sceptical about the Asus ROG Strix XG129C, this isn't a proper monitor. It's a device for keeping track of, say, your PC's operating status or providing space for some supplementary application controls. Somehow, I ended up quite liking it. And not just because I'm a total sucker for screens. But more on my journey in a moment. First, let's cover off the core speeds and feeds.

The 24:9 aspect translates into a native resolution of 1,920 by 720 pixels and a pixel density of 166 DPI, the latter being about the same as a 27-inch 4K monitor. So, it's pretty crisp.

It uses IPS panel tech and tops out at 300 nits for brightness and 75 Hz in terms of refresh rate. Oh, and quite importantly, it supports capacitive 10-point multi-touch input. For the record, the touch functionality works essentially out-of-the-box provided you are connecting via USB-C.

Speaking of which, there are two USB-C ports. Both will drive the display, while one doubles as power in and the other supports 20 W of power out. So you can, for instance, connect via a single USB-C cable to both drive the display and power it, which is the ideal in terms of cable management.

That's easy from a laptop, but less of a given with a desktop PC. There are desktop graphics cards with USB-C out that support up to 30 W of power delivery and so would be able to power the display and drive it. But that setup would not support the touch functionality, which requires data passthrough.

So, for a true single-cable solution on a desktop, it all gets a bit complicated. In theory, you could do it with a motherboard that supports Thunderbolt and DisplayPort passthrough from the iGPU and thus have the Asus ROG Strix XG129C running from the iGPU in your CPU and your main monitor running off your graphics card.

I understand this setup is supported, for instance, on Intel Z890 motherboards, again in theory. In practice, results seem to vary and there can be issues with the Thunderbolt port supplying sufficient power. In a worst case scenario, then, you might need to use the HDMI port for the video signal, one of the USB-C ports to power the display and another to send touch data to your PC, which is pretty messy.

But I think a decent chunk of users will be able to get it working with two of the three ports hooked up. If one of those is the HDMI port, you'll then have one cable coming out of the side, one out of the top. It would have been tidier if Asus had located all the ports on one side. Never mind.

That ergonomic glitch aside, this is otherwise a really nicely built display. It feels high quality and the screw-in kick stand is very sturdy and capped with a rubberised foot to provide grip, as is the base of the display itself.

You also get a proper OSD with a joystick controller and full set of features, just like a conventional Asus monitor. What's more, you can control the OSD with the touchscreen input, which is a nice, well, touch.

If that's the features covered, what is the Asus ROG Strix XG129C like in use? Asus bundles a year's subscription to an OEM version of Aida 64, so you've got something to play with immediately.

That's handy, because it would be a nightmare to configure the standard version of Aida 64 to fit this display perfectly. As it is, you can use it to track CPU and GPU operating speeds and temps, storage use, that kind of thing. You'd have thought a permanent license would be part of the deal, but at least you have something to run immediately.

Outside of that, your mileage will vary depending on your preferences. You could use this display to do anything from keep track of social media feeds or work collab' platforms like Slack while you game, to using it to situate secondary controls or a timeline for video editing.

Much will depend on the exact applications you use and what kind of tools and controls they offer, how much configurability they support and so on. But what I would say is that, for this display to really deliver, those tools and controls need to be touch friendly.

That's generally quite a big ask, given that so many applications are either poorly or not at all optimised for touch. Ironically, that includes Asus' own Display Widget Controller app, which really doesn't work at all well on this display. Whatever, it's the ability to use the touch interface, as opposed to needing to "mouse over," and the immediacy that offers, that is most compelling.

Of course, that in itself can create problems with application focus, in-game. And there's no doubting the Asus RoG Strix XG129C works most reliably and consistently as a status display rather than something you directly interact with. But the shortcomings there are largely down to how Windows operates, rather than the display itself.

In the end, this little display is very much in the nice-to-have rather than must-have category. It's also the sort of device where your benefit from and use of it will depend on a whole range of factors.

Some people may be able to get quite a bit of value out of it as a touch device for controlling application tools and menus, such as video, image and music editing. But I would advise caution in that regard. You'll need to do a little homework to see what really works. Don't assume too much.

On the other hand, if you just like the idea of a small secondary display for keeping track of your PC's hardware status or viewing work and social feeds, you can be more confident that the Asus ROG Strix XG129C will deliver.

What's more, as someone who prefers a single ultrawide monitor to a pair of more conventional 16:9 panels, what this mini monitor offers is a way to have the status-tracking functionality that a second screen offers, but keep it small and ergonomically tidy. I don't need a huge screen to watch a bit of Slack or keep track of emails arriving in the background, in other words. The Asus ROG Strix XG129C does that kind of job rather nicely.

Which brings us to the final important factor, namely price. At $199, the Asus ROG Strix XG129C is either pretty reasonable for such a specialised device or crazy money given what 200 bucks buys you these days in the monitor market. This year during Amazon Prime week, I found some KTC 34-inch ultrawides for barely over $200, while $250 will get you a 27-inch 4K dual-mode panel and $150 a decent 1440p panel. It's all a bit bananas.

In that context, the Asus ROG Strix XG129C is definitely a luxury item rather than a value proposition. You'll certainly have to specifically want a smaller, and in some ways less functional, screen to choose it over a 27-inch 1440p panel for less money.

But, personally, I actually might. I already have a large ultrawide for my primary viewing. I don't want another full-sized screen hanging miles off to the side. This nifty little accessory display can sit below my main monitor and would be just the thing to have running discretely, keeping me up to date. I've just got to decide if that's worth $200 to me…

If you're in the market for a pricey, studio-like USB microphone for high-quality audio capture, you've probably considered one of Shure's offerings. You'd do well to look at the Shure MV7+, a dynamic podcast mic that's cemented itself into our best microphone for gaming guide for some time now. Razer, however, has come up with something that might not beat the MV7+ outright, but certainly gives it something to think about.

It's called the Razer Seiren V3 Pro, and it retails for $250/£250. That's around $40-45 cheaper than the MV7+ at current prices, and while the V3 Pro isn't quite able to deliver the same ultra-refined, smooth vocal capture, it's got a lot going for it in terms of sheer flexibility.

It's more usable in a desktop position, for a start. You also get access to an AI noise reduction feature, a noise gate, a compressor, and a parametric EQ with plenty of well-tuned presets. There's also an auto setup feature, which I had some real trouble with. I'd say it was better ignored, or if needed, used to provide a rough baseline you can tweak for yourself later.

Still, many great vocal tones can be pulled from this particular unit—albeit with a touch of fiddling. It's got a tendency towards a detailed capture, but still reveals the odd rough edge that takes a little tweaking to filter out. But with the flexibility on offer here, the V3 Pro makes a good case for itself as a versatile vocal mic.

It's not quite on the Shure's level when it comes to finding that podcast-like, streaming friendly, studio-grade sound. But it's a jack-of-all-trades with a lot to offer, provided you're prepared to put a little time into it.

Razer Seiren V3 Pro features

The Razer Seiren V3 Pro is a dynamic microphone, which means it's not as sensitive as a traditional desktop condenser mic. You'll need to move some air in the near vicinity to get the most out of it, although unlike the Shure MV7+, Razer has included a screw-in stand base to allow it to function as a desktop mic straight from the box.

It makes use of a 30 mm dynamic capsule with a frequency response of 50 - 16,000 Hz, and supports audio capture up to a 32-bit/96 kHz resolution, along with 32-bit float support through a Razer Synapse setting. This format technically allows for higher peaks without clipping or distortion.

Do you need 32-bit float support? Probably not. But Razer says it's perfect for streamers who speak quietly, and then suddenly yell at… I don't know, something jumping out of a cupboard in a horror game.

The Seiren V3 Pro is a dual-connectivity mic, with an XLR socket and USB Type-C port integrated into the bottom of the chassis, although you can't use both at once. In terms of controls, you get a manual gain control wheel on the bottom, and a large tap-to-mute button on the top.

Around the bottom of the foam pop filter is a tastefully small ring of RGB Chroma lighting, and the microphone itself is attached to an adjustable one-sided arm with a ⅝-inch boom arm socket in the bottom.

It's a good-looking, professional-feeling object, with a chunky, overbuilt design that underscores its "Pro" moniker. If I had one ergonomic critique, I'd say the pop filter is a touch loose in terms of fit, which means it has a tendency to slip slightly out of position at certain angles. Mind you, some Shure mics have the same problem, and I've not had it fall off entirely under testing, so it's not a huge deal.

Razer Seiren V3 Pro sound

I'll go more into the Seiren V3 Pro's auto setup option in the software section below, but as a quick note: It's far from perfect. However, the microphone itself is a well-equipped unit, and the manual software options work fairly well. The clips below were recorded with custom settings, rather than auto setup:

Desktop position

In a desktop position, it's also rather quiet—which is to be expected from a dynamic microphone. I had to boost the gain settings to record the clip above, with a dose of compression and some AI noise cancelling to even out the volume and remove some unwanted nasties. The noise gate is a touch aggressive, which means I had to set it fairly low and allow some background hiss to remain in the final desktop-position take.

Close-up position

Still, Razer advertises this mic as delivering "studio sound", and while it takes a little fiddling in the software to get there, the close-up results sound very good indeed. The desktop position audio is also perfectly decent, if a touch noisy. It can do both, but being a dynamic microphone, it still likes being close to your lips for those toasty tones.

I'd say what we're dealing with here is a good-quality capsule that leans towards detail, with the EQ presets and software features richening things up. There's nothing wrong with that, but the Shure MV7+ sounds like a higher-quality unit from the get-go.

Razer Seiren V3 Pro software

Razer has made a pretty big deal of this mic's auto-setup feature, so I was curious to see what it made of my quiet living room in a desktop position. Upon connecting the mic for the first time, Synapse asks you to speak normally into the microphone to record a test clip, before taking a sample of the room ambience to adjust the noise-cancelling features.

It then asks you to pick a primary purpose, such as streaming, podcasting, gaming capture, etc, before adjusting all of its own settings.

Which, in my case, led to an overly-aggressive noise gate that clipped my voice so badly that every other word was missed out in my first test clip. Oh dear. Trying again, it seems that the auto setup gets very confused with my particular use case, with the mic placed 10 -15 cm away from my mouth next to a window with occasional traffic.

Desktop position, auto settings

Despite having a very good compressor, AI noise suppression, and the aforementioned noise gate to play with, the auto setup struggled to find a good gain level, set the gate far too aggressively, and chose to ignore the compressor and AI noise suppression sliders entirely. Which again resulted in an uneven sound with plenty of words being cut off.

Close-up, auto settings

Still, with a bit of tweaking (essentially, I fiddled with the noise gate settings until it stopped cutting off the start of my words, gave it a gain boost, and compensated with AI noise suppression and a touch of compression), the Razer captures a rich vocal tone overall, as you can hear in the Sound section above.

However, the auto setup feature regularly makes mistakes in terms of its settings choices, which is likely to leave many users scratching their heads. It's a nice idea on paper, but in practice, I'd leave it well alone.

Top marks, however, go to the parametric EQ presets. I leapt straight to the "podcast" setting and was presented with an EQ curve that looked very familiar. I've done some podcast audio engineering in a previous life, and Razer's EQ curve looks… well, it looks very much like one of mine. Which is obviously the best podcast EQ you'll ever find, of course.

You can tweak the EQ around to your heart's content, but each of Razer's default options is a pretty good start. I've noticed a tendency to boost the bass frequencies in each preset with a sharp drop-off in terms of sub-bass, which is often good practice when hunting for a 'rich' sound. Still, as noted above, it's the EQ that's doing a lot of the work on what is otherwise a good, but not mind-blowing capsule.

But overall? The software design is fairly straightforward, as long as you're prepared to do some of the work yourself. I wish the auto setup was better when it comes to some of its settings choices, but there are enough options here to easily correct its mistakes.

Razer Seiren V3 Pro value

At $250, the Razer Seiren V3 Pro is far from cheap. However, it is cheap-er than the Shure MV7+, and finds itself shooting for the same target market. Namely, those who want a near-studio-level microphone for semi-pro streaming audio.

The problem is, the Shure microphone is simply better at performing that goal, at least in a close-up position. While it's a chonkier unit and less versatile, its key selling point is the plug-and-play nature of both its software and its internal hardware.

I've spent half a lifetime screwing around with audio settings, and the Shure MV7+ is so darn good at finding a professional sound with the minimum of tweaks, I consider it the be-all and end-all of semi-pro vocal mics. I had hoped that the Razer's auto-setup feature would close that gap, but unfortunately it had me hunting through the settings at multiple points during testing.

That being said, the Seiren V3 Pro is a very capable microphone once you've tweaked it for yourself. It also looks great and has a chassis design that oozes refinement. And, despite its flaws, I think its sonic characteristics (after the requisite effort) make it worth considering. Particularly if you plan on mixing things up between desktop and close-up positions.

Just leave the auto setup well alone, yes?

The Dark Rock Pro 6 and Dark Rock 6 are a huge upgrade over Be Quiet's previous air coolers. Where it had previously struggled to keep up with our picks for the best air coolers, these new arrivals are amongst the top performers in my testing.

Be Quiet hasn't lost its eye for aesthetics, either. The all-black design of the Dark Rock 6 duo looks even more impressive out of the box than previous models. In that sense, Be Quiet has something of a unique selling point here: it's turned an imposing mass of aluminium into a pretty thing.

Yet there's an elephant in the room. Despite my appreciation for the upgraded performance, aesthetics, and easy installation, I just can't look past the price attached to these two coolers. Both are very expensive, at $110/$130 for the standard/Pro models, respectively. That's one helluva markup on some of the excellent air coolers we rank highly today, an added cost of around $60–$100. When air coolers are seen as the more affordable alternative, these prices just don't add up.

Specifications

The Dark Rock 6 comes with a Silent Wings 135 mm fan. That's a whole 15 mm larger than the one found on this cooler's predecessor, the Dark Rock 5. That's one clear benefit of the newer model: a big fan can push more air at lower RPM compared to a small fan. The 135 mm fan used here can spin up to 2000 rpm.

This is also a "custom-designed" fan for this cooler, rather than the off-the-shelf Silent Wings 4 found on the previous model. That said, from what I can gather, and I only have a smaller Silent Wings 4 for comparison, the main difference between this custom-designed fan and the Silent Wings 4 is a slightly smaller hub. I have measured the hub on both and they are roughly 5 mm apart—so that'd explain the 135 mm sizing on the newer custom models.

The Dark Rock 6 also comes with a semi-passive mode. There's a small switch hidden on the side of the fan shroud, marked 'P' and 'Q'. No prizes for what these both mean.

• Performance mode: runs the fans based on whatever you've set on your motherboard header—the cooler's single fan uses a 4-pin PWM connection.
• Quiet mode: overrides the fan header to reduce RPM at certain duty cycles, and switches it off entirely below 40%.

The Dark Rock Pro 6 adds another 120 mm fan into the mix. The two fans are connected together with a 7-pin connector. It also includes one extra heatpipe compared to the standard model, seven to six, which makes it that bit more capable, as you'll see in the thermals section below.

Thermals

The Dark Rock Pro 6 and Dark Rock 6 are very capable air coolers. Both manage to keep a Core i7 14700K at a reasonable temperature under load. The Dark Rock 6 generally performs similarly to the Arctic Freezer 36, if at times even outperforming our pick for the best budget air cooler, except at higher power draw. Meanwhile, the Dark Rock Pro 6 performs very well compared to the rest, if only falling short of the mighty Thermalright Peerless Assassin.

Thermal performance

The chart below shows the delta over ambient, accounting for any fluctuations in ambient temperature between tests

Baldur's Gate 3 @ 1.275 V | ΔTx264 @ 1.275 V | ΔTPrime95 @ 1.275 V | ΔTPrime95 @ 1.325 V | ΔTAmbient temperature

Max ΔT

Avg ΔT

Be Quiet! Dark Rock Pro 6

34

28

Be Quiet! Dark Rock 6

34

30

Be Quiet! Dark Rock 5

37

33

Arctic Freezer 36

34

31

Arctic Freezer 36-S A-RGB

34

31

Thermalright Peerless Assassin 120 SE

32

28

010203040

Baldur's Gate 3 @ 1.275 V | ΔT Data

Product	Value
Be Quiet! Dark Rock Pro 6	34 Max ΔT, 28 Avg ΔT
Be Quiet! Dark Rock 6	34 Max ΔT, 30 Avg ΔT
Be Quiet! Dark Rock 5	37 Max ΔT, 33 Avg ΔT
Arctic Freezer 36	34 Max ΔT, 31 Avg ΔT
Arctic Freezer 36-S A-RGB	34 Max ΔT, 31 Avg ΔT
Thermalright Peerless Assassin 120 SE	32 Max ΔT, 28 Avg ΔT

Max ΔT

Avg ΔT

Be Quiet! Dark Rock Pro 6

43

33

Be Quiet! Dark Rock 6

43

34

Be Quiet! Dark Rock 5

48

38

Arctic Freezer 36

43

35

Arctic Freezer 36-S A-RGB

44

36

Thermalright Peerless Assassin 120 SE

40

32

015304560

x264 @ 1.275 V | ΔT Data

Product	Value
Be Quiet! Dark Rock Pro 6	43 Max ΔT, 33 Avg ΔT
Be Quiet! Dark Rock 6	43 Max ΔT, 34 Avg ΔT
Be Quiet! Dark Rock 5	48 Max ΔT, 38 Avg ΔT
Arctic Freezer 36	43 Max ΔT, 35 Avg ΔT
Arctic Freezer 36-S A-RGB	44 Max ΔT, 36 Avg ΔT
Thermalright Peerless Assassin 120 SE	40 Max ΔT, 32 Avg ΔT

Max ΔT

Avg ΔT

Be Quiet! Dark Rock Pro 6

43

41

Be Quiet! Dark Rock 6

43

39

Be Quiet! Dark Rock 5

50

47

Arctic Freezer 36

44

40

Arctic Freezer 36-S A-RGB

46

42

Thermalright Peerless Assassin 120 SE

44

38

015304560

Prime95 @ 1.275 V | ΔT Data

Product	Value
Be Quiet! Dark Rock Pro 6	43 Max ΔT, 41 Avg ΔT
Be Quiet! Dark Rock 6	43 Max ΔT, 39 Avg ΔT
Be Quiet! Dark Rock 5	50 Max ΔT, 47 Avg ΔT
Arctic Freezer 36	44 Max ΔT, 40 Avg ΔT
Arctic Freezer 36-S A-RGB	46 Max ΔT, 42 Avg ΔT
Thermalright Peerless Assassin 120 SE	44 Max ΔT, 38 Avg ΔT

Max ΔT

Avg ΔT

Be Quiet! Dark Rock Pro 6

62

54

Be Quiet! Dark Rock 6

63

58

Be Quiet! Dark Rock 5

70

65

Arctic Freezer 36

59

55

Arctic Freezer 36-S A-RGB

59

55

Thermalright Peerless Assassin 120 SE

60

51

020406080

Prime95 @ 1.325 V | ΔT Data

Product	Value
Be Quiet! Dark Rock Pro 6	62 Max ΔT, 54 Avg ΔT
Be Quiet! Dark Rock 6	63 Max ΔT, 58 Avg ΔT
Be Quiet! Dark Rock 5	70 Max ΔT, 65 Avg ΔT
Arctic Freezer 36	59 Max ΔT, 55 Avg ΔT
Arctic Freezer 36-S A-RGB	59 Max ΔT, 55 Avg ΔT
Thermalright Peerless Assassin 120 SE	60 Max ΔT, 51 Avg ΔT

Be Quiet! Dark Rock Pro 6

20

Be Quiet! Dark Rock 6

20

Be Quiet! Dark Rock 5

21

Arctic Freezer 36

27

Arctic Freezer 36-S A-RGB

23

Thermalright Peerless Assassin 120 SE

22

07.51522.530

Ambient temperature Data

Product	Value
Be Quiet! Dark Rock Pro 6	20
Be Quiet! Dark Rock 6	20
Be Quiet! Dark Rock 5	21
Arctic Freezer 36	27
Arctic Freezer 36-S A-RGB	23
Thermalright Peerless Assassin 120 SE	22

They also both easily outperform the Dark Rock 5 (non-Pro) in my testing. That was, by far, the worst performer of the lot in my latest batch of testing, which shows just how significant a leap this new lineup has been by way of performance for Be Quiet.