AMD Ryzen AI Max "Strix Halo" APU Reviews Reportedly Arriving Imminently

[T0@st]

Yesterday, the ASUS Chinese office announced a special event—on February 25—dedicated to launching a next-gen AMD APU-powered premium notebook model: "ROG Magic X (or Illusion X) is the first to be equipped with the Ryzen AI MAX+ three-in-one chip, which can efficiently coordinate multiple modes, provide combat power and computing power on demand, and can handle e-sports, creation and AI with one chip!" The manufacturer's Weibo post has generated plenty of buzz; industry insiders reckon that reviews could be published today (February 18)—HXL/9550pro informed VideoCardz with a not so cryptic message: "STX-Halo NDA: Feb 18th 2025."

Western press outlets point out that the ROG Magic X is a local variant of the familiar ROG Flow Z13 design; a 2025 refresh brings in AMD's much anticipated "Strix Halo" APU design. Team Red-authored marketing material and pre-release evaluation leaks have hinted about impressive integrated graphics solution performance; equalling or even exceeding that of previous-gen dGPUs. Well-known North American hardware review outlets have dropped hints (NDA permitting) about AMD's Ryzen Al Max+ 395 and Max 390 processors. Hardware Canucks could barely contain their excitement regarding the potent Zen 4 and RDNA 3.5 combo package; to the point of wish listing a potential direct successor: "Strix Halo is one of the most exciting things launched into the PC space in the last half decade. Full stop...AMD can't keep this as a one-off. If it's followed up with Zen 6 and RDNA 4 next year...watch out." Naturally, Team Red's cutting-edge mobile CPU technology is arriving in devices with high asking prices. The aforementioned ROG Flow Z13 2025 model—configured with top specs—is priced at $2699. Notebookcheck reckons that ASUS has tacked on an extra $500, since an announcement of initial pricing at CES 2025.



View at TechPowerUp Main Site | Source

 

[Daven]

Flowery language aside Strix Halo is actually a really big deal. While Apple silicon is locked to MacOS and ARM ISA, this is the first attempt in a long while of having such an SoC in the Windows/x86 space. Strix Halo basically brings the Xbox and PlayStation architecture philosophy to laptops.

 

[Sound_Card]

It has discrete 4060 / 7600 performance, which is mad wild.

 

[ThomasK]

Hopefully we'll see models with the option to max out the TDP at 120w.

 

[R0H1T]

Already here ~

https://www.notebookcheck.net/Asus-ROG-Flow-Z13-GZ302EA-Convertible-Review-AMD-s-Strix-Halo-GPU-is-neck-and-neck-with-the-RTX-4070-Laptop.963266.0.html

Hopefully we'll see models with the option to max out the TDP at 120w.

Patience 0'x3d one

 

[HD64G]

 

[hsew]

Starting price of $1999 is a bit high (I'm willing to blame ASUS for that), but the fact that they list a 128GB sku is promising.

ASUS' Store link

ROG Flow Z13 (2025) GZ302 | 2-in-1 Gaming Tablet | ROG USA

Game and create on the go with the 2025 ROG Flow Z13: experience the power of a full-fledged gaming laptop packed in a 13" tablet with a long-lasting battery.

rog.asus.com

 

[Daven]

Looks like 3D stacked cache is also possible with these chips according to Videocardz. The TSVs are present.

AMD Ryzen AI MAX 300 "Strix Halo" reviews are here, dawn of mid-range discrete GPUs - VideoCardz.com

 

[Chaitanya]

Looking forward to HP options(for their much better after sales service than Shitsus) with 128GB RAM options when the AMD exclusive partnership ends.

Edit: looks like atleast Shitsus hasnt castrated the microSD slot by going with UHS-I interface and its a UHS-II interface capable of transfer speeds of around 250MBps.

 

[R0H1T]

In a bigger chassis this would definitely perform better, the z13 would be holding it back at least to a fair(small?) extent.

 

[JustBenching]

So its between a 4060m and a 4070m. That's nuts. The problem with these dGPUs is that they need ~30-35w at the very minimum to function, if the iGPU can actually drop to very low power levels and still retain good performance - then well, mobile gpus are dead for laptops.

 

[igormp]

So its between a 4060m and a 4070m. That's nuts. The problem with these dGPUs is that they need ~30-35w at the very minimum to function, if the iGPU can actually drop to very low power levels and still retain good performance - then well, mobile gpus are dead for laptops.

Idle power is still awful compared to other iGPU laptops, but way better than other dGPU laptops, so it's a weird in-between.

It's also way worse than the M4 Pro when it comes to power efficiency (and even in raw performance in many cases).

 

[TPUnique]

Excited to see that the performance leaks are now confirmed, a game changer indeed.

Now this needs to be released on an ITX motherboard, which Minisforum will prolly do by the middle of this year.

And in the future, AMD needs to work on dramatically improving the chip's bandwidth : ideally, two generations down the line, their APU should reach parity with Apple's M2 Ultra (800 GB/sec).

 

[R0H1T]

And in the future, AMD needs to work on dramatically improving the chip's bandwidth : ideally, two generations down the line, their APU should reach parity with Apple's M2 Ultra (800 GB/sec).

Not going to happen unless they go x3d &/or HBM, there were rumors last year that Sammy(SK hynix?) was working on mobile HBM which is the only thing that could come close to LPDDRx level efficiency ~

Mobile HBM: Empowering the Next Generation of Mobile Device Memory | ABNewswire

It's also way worse than the M4 Pro when it comes to power efficiency (and even in raw performance in many cases).

Power efficiency is hard to gauge given all the crud that runs in the background. Can someone for once run an absolute "barebone" Windows installation without useless programs/services eating up all that battery

 

[Denver]

It's odd that Asus sets the "VRAM" to 4GB by default, especially since HC shows this impacts performance.

Strix Halo's much more efficient than I expected. I thought it would spike above 200W, but it's staying around the well-defined TDP limits.

 

[TPUnique]

Not going to happen unless they go x3d &/or HBM, there were rumors last year that Sammy(SK hynix?) was working on mobile HBM which is the only thing that could come close to LPDDRx level efficiency ~

Mobile HBM: Empowering the Next Generation of Mobile Device Memory | ABNewswire

As far as I know, Apple uses standard LPDDR in its M2 Ultra chip.

 

[R0H1T]

Standard LPDDR like? I used "x" to denote any LPDDR be it LPDDR4/x or LDPPR5/x because they're all pretty close to each other efficiency wise.

 

[Daven]

Standard LPDDR like? I used "x" to denote any LPDDR bet it LPDDR4/x or LDPPR5/x because they're all pretty close to each other efficiency wise.

As far as I know, Apple uses 'regular' DDR memory with up to 8 channels (512-bit). That's why the bandwidth is so high. I see no reason AMD cannot do this on a future chip unless you can point to some fundamental limit that prevents 8-channel operation.

 

[R0H1T]

It's LPDDR5 running at 6400 MT/s & the M2 ultra is practically an MCM.

 

[TPUnique]

Standard LPDDR like? I used "x" to denote any LPDDR be it LPDDR4/x or LDPPR5/x because they're all pretty close to each other efficiency wise.

I have a hard time finding a specific info, but it's prolly LPDDR5X.
In any case, it doesn't seem to be HBM.

So if Apple could reach that sort of bandwidth two years ago in 2023, I don't see a reason why AMD couldn't do the same in 2027. And at a more reasonable price.

 

[R0H1T]

It's on Wikipedia as well & I watched the M2 ultra's launch/reveal ~

Industry-Leading UltraFusion Technology
M2 Ultra is built from two M2 Max dies connected through UltraFusion, Apple’s industry-leading, custom-built packaging technology. UltraFusion uses a silicon interposer that connects the dies with more than 10,000 signals, providing over 2.5TB/s of low-latency interprocessor bandwidth.
UltraFusion’s architecture enables M2 Ultra to appear as a single chip to software. This means code doesn’t need to be rewritten to utilize the extreme performance of M2 Ultra and makes UltraFusion unlike anything else in the industry.

Apple introduces M2 Ultra

Apple today announced M2 Ultra, a new system on a chip (SoC) that delivers huge performance increases to the Mac and completes the M2 family.

www.apple.com

They're not making any standard DDR5 based chips anymore, that'd defeat the whole purpose of making the Mx chips.

 

[Sound_Card]

Medusa Point looks even more mad wild with Zen 6 and RDNA 4(.5?). 2026.
I'm far more interested in 'console mini PC's' these days than building a mid tower. I think a lot of people are focused on this market.

 

[L|NK|N]

Medusa Point looks even more mad wild with Zen 6 and RDNA 4(.5?). 2026.
I'm far more interested in 'console mini PC's' these days than building a mid tower. I think a lot of people are focused on this market.

Same! I would really love for this and for Valve to step up a re-boot of the Steam Machine / SteamOS in the desktop space, now that the Steam Deck OS has come so very far.

 

[HD64G]

Very close to 4070 mobile performance at half the power draw if you combine the CPU's and the 4070's seperate wattage. That means that if the APU's power limit is increased to AMD's max 120W, it will easily beat 4070 mobile. Only disadvantage is scarcity that allows for ultra high pricing. I think that will get better in 4-5 months down the road once more AIBs make those and cut-down iterations arrive. Other than that, a trully innovative product for the home computing market. Small factor or mini-PCs will become complete with those fabulous APUs that can serve any-purpose any-time.

 

[mama]

Looks like the perfect answer to a quick and dirty PC with 4070 laptop performance.