AMD's next-gen 10-core 'Medusa Point' APU shows up on Geekbench again, with its best score yet — leaked SKU outpaces every other x86 mobile chip in the single-core test

AMD is expected to refresh both its desktop and mobile lineups with Zen 6 next year. Ryzen AI 500 series, codenamed Medusa Point, will be the company's next-gen mobile family and we've already seen a 10-core part leak from it a couple of times. Now, that same SKU is back with another Geekbench listing, this time posting its best results that put it ahead of any mobile chip currently made by Intel or AMD (in one way).

Medusa Point 10-core SKU listed on Geekbench — (Image credit: Future)

AMD's upcoming chip ended up scoring 3,329 points in the single-core test and 16,555 points in the multi-core test. The Geekbench page once again shows the platform name as "AMD Plum-MDS1," which we've known to be associated with Medusa Point for a while now. The SKU is listed as "AMD Eng Sample 100-000001713-33_N," which is identical to how we saw in the last leak, basically confirming we're looking at the same chip.

The single-core result is almost 5% higher while the multi-core score is about 9.7% higher compared to the numbers we saw previously. In fact, the single-core result is better than any mobile part on Geekbench apart from the new Snapdragon X2 Elite lineup. Compared to desktop equivalents, it's very close to the 9900X and the 9800X3D — two absolute desktop powerhouses carrying 12 cores and 8 cores, respectively.

Latest Videos FromTom's Hardware

Watch full video here:

Swipe to scroll horizontally

SKU	Config	Single-Core Score	Difference
Qualcomm X2 Elite X2E-90-100	Mobile, 18C	3,573	100%
AMD Ryzen 9 9900X	Desktop, 12C	3,333	93%
AMD Ryzen 7 9800X3D	Desktop, 8C	3,333	93%
AMD Ryzen 9 565?	Mobile, 10C	3,329	93%
AMD Ryzen 5 9600X	Desktop, 6C	3,318	93%
Intel Core i9-14900KS	Desktop, 24C	3,226	90%
Intel Core Ultra 9 285K	Desktop, 24C	3,195	89%
AMD Ryzen 9 565? (Previous)	Mobile, 10C	3,174	89%

The multi-core score puts this 10-core SKU right alongside the Ryzen AI Max+ 390, which is just a tier below the top-end AI Max+ 395 from Strix Halo. Since this is a 10-core part, we can infer it's the Ryzen AI 9 565 because the Ryzen AI 9 365 (from Strix Point) and Ryzen AI 9 465 (from Gorgon Point) share the same configs. The Ryzen AI 9 365 has an average score of 12,410, so Medusa Point's offering is 33% faster in this regard.

Swipe to scroll horizontally

SKU	Config	Multi-Core Score	Difference
Qualcomm X2 Elite X2E-90-100	Mobile, 18C	18,026	78%
AMD Ryzen 9 9950X3D	Desktop, 16C	22,169	96%
AMD Ryzen 9 9900X	Desktop, 12C	19,696	86%
AMD Ryzen 7 9800X3D	Desktop, 8C	18,317	80%
AMD Ryzen 9 565?	Mobile, 10C	16,555	72%
AMD Ryzen 5 9600X	Desktop, 6C	14,742	64%
Intel Core i9-14900KS	Desktop, 24C	23,008	100%
Intel Core Ultra 9 285K	Desktop, 24C	22,472	98%
AMD Ryzen 9 565? (Previous)	Mobile, 10C	15,092	66%

The Geekbench listing also reports the correct clock speeds this time — 2.0 GHz is still lasted as the base frequency, but the boost frequency is marked at 5.37 GHz, which makes a lot more sense. Moreover, the listing shows 10MB of L2 cache and 32MB of L3 cache while the Ryzen AI 9 465 had a combined cache tally of 34MB.

Overall, this benchmark is an incredibly promising showing for Medusa Point. The Red Team's next-gen mobile lineup seems to benefit heavily from an architectural jump thanks to the Zen 6 silicon. Keep in mind that this is the company's mainstream mobile offering that's supposed to compete with Intel's Panther Lake, while Nova Lake mobile and AMD's own Gator Range (Ryzen 10000) will serve only as high-end CPUs.

Follow Tom's Hardware on Google News, or add us as a preferred source, to get our latest news, analysis, & reviews in your feeds.

TOPICS

Hassam Nasir is a die-hard hardware enthusiast with years of experience as a tech editor and writer, focusing on detailed CPU comparisons and general hardware news. When he’s not working, you’ll find him bending tubes for his ever-evolving custom water-loop gaming rig or benchmarking the latest CPUs and GPUs just for fun.

8 Comments Comment from the forums

usertests

2.0 GHz is still lasted as the base frequency
listed

Note: 10-core Medusa Point is considered to be an 8-core Krackan Point successor.

We're looking at gains with Zen 6 that are similar to Zen 4, from a decent IPC uplift and higher clocks. Apparently up to 5.37 GHz turbo on this part, only a little higher than the 5.1-5.2 GHz seen this generation from Krackan/Strix/Gorgon, but I think sustained all-core clocks are likely doing much better due to efficiency improvements.

Is "cluster" analogous to CCX in Geekbench listings? I think it's probably one 10-core CCX here. If 32 MiB L3 cache is accurate that will be a doubling from Krackan, very nice.

I guess this SKU will have 8 CUs of RDNA4m, but no confirmation seen anywhere yet. Strix Point was often bandwidth starved, so 8 CUs of RDNA4m may perform much better than you would expect, also with full fat/speed FSR4 support.

Krackan looked nice to me when it was available in $400-500 laptops (not anymore, but I see a good recent but expired deal). Medusa Point will be a great improvement, softening the blow of the terrible pricing.
Reply
thestryker

I've been thinking about the cache configurations since the first leaks as 32MB doesn't really make sense. My assumption has been that the regular mobile cores may just have half of what desktop does leading to 24MB. Historically the C cores have halved regular cores which would drop that to 12MB. Obviously 24 + 12 isn't 32 so that would mean one of two things to me: the C cores are sharing 8MB or AMD has improved the way they're disabling cache and are doing it in 4MB chunks instead of the 8MB with Zen 4/5.

As for the performance I'd point out that the 9600X has presumably the same boost clock as this sample being tested.
Reply
usertests

thestryker said:
I've been thinking about the cache configurations since the first leaks as 32MB doesn't really make sense. My assumption has been that the regular mobile cores may just have half of what desktop does leading to 24MB. Historically the C cores have halved regular cores which would drop that to 12MB. Obviously 24 + 12 isn't 32 so that would mean one of two things to me: the C cores are sharing 8MB or AMD has improved the way they're disabling cache and are doing it in 4MB chunks instead of the 8MB with Zen 4/5.
We should be looking at 4x Zen 6 + 4x Zen 6c + 2x Zen 6 LP. If you extrapolate from Krackan and assume a 50% across the board increase in L3 cache for Zen 6 products, you get to 24 MiB. That's ignoring the two LP cores. We have no examples yet of how AMD will treat LP cores, but I guess you can assume zero L3 for now. Maybe these LP cores are in a separate CCX, but it's a monolithic die and we will see 12-core CCXs this generation.

32 MiB unified would be a crazy improvement when Krackan had 16 and Strix Point had the 16 + 8 split. It's safe to assume misdetection for now.
Reply
thestryker

usertests said:
32 MiB unified would be a crazy improvement when Krackan had 16 and Strix Point had the 16 + 8 split.
I haven't seen any shots of Krackan showing the core design so I don't know if it's similar to Strix or Phoenix. With Phoenix the regular and C cores directly shared the cache so I suppose that would be another option where it's just a single shared cache rather than split.

usertests said:
Maybe these LP cores are in a separate CCX
It makes sense that they'd be separated in some fashion as that's what Intel has done with all of theirs even when on the same silicon.
Reply
usertests

usertests said:
If you extrapolate from Krackan and assume a 50% across the board increase in L3 cache for Zen 6 products, you get to 24 MiB.
That might not be a great assumption since budget APUs should be on an N3 node while desktop CCDs are on an N2 node with denser SRAM. But who knows.
Reply
Thunder64

usertests said:
We should be looking at 4x Zen 6 + 4x Zen 6c + 2x Zen 6 LP. If you extrapolate from Krackan and assume a 50% across the board increase in L3 cache for Zen 6 products, you get to 24 MiB. That's ignoring the two LP cores. We have no examples yet of how AMD will treat LP cores, but I guess you can assume zero L3 for now. Maybe these LP cores are in a separate CCX, but it's a monolithic die and we will see 12-core CCXs this generation.

32 MiB unified would be a crazy improvement when Krackan had 16 and Strix Point had the 16 + 8 split. It's safe to assume misdetection for now.

I thought the rumor was the LP cores were on the IOD? That seems to make sense.
Reply
usertests

Thunder64 said:
I thought the rumor was the LP cores were on the IOD? That seems to make sense.
This product should be monolithic, and even for the Zen 6 APUs that are not monolithic, they supposedly have an IOD/central chiplet that has all 3 types of cores and the iGPU. They can optionally have a desktop chiplet added to boost the core count.

It's the desktop CPUs that have physical separation of LP cores on the IOD, but only if AMD commits to it. We've also seen a leaker claim the iGPU will be dropped in favor of an NPU! So Olympic Ridge details are up in the air for now.
Reply
JamesJones44

In fact, the single-core result is better than any mobile part on Geekbench apart from the new Snapdragon X2 Elite lineup
Also apart from the Apple Silicon lineup. The base M5 has a single core score of 4223 and multi-core score of 17475
Reply

Show more comments