AMD does not have a new desktop CPU generation this year and is just refreshing previous Zen 5 chips, as in the Ryzen AI 400 \u201cGorgon Point\u201d<\/a> APUs announced at CES 2026. While a matching refresh of last year’s high-performance Ryzen AI Max 300 processors<\/a> was not announced there, it appears this lineup will also receive similar a minor update. That will have to suffice for a while, as a brand-new generation is still far away.<\/p>\n Last week, the website VideoCardz reported that, according to its sources, AMD is indeed preparing a Strix Halo refresh, which is the first generation of Ryzen AI Max \u201cHalo\u201d line of high-end APUs, which use chiplets and 256-bit LPDDR5X memory, enabling configurations with up to 16 Zen 5 cores and an unusually powerful integrated GPU with 40 CUs (2560 shaders) based on the RDNA 3.5 architecture.<\/p>\n The refresh that is in preparation is codenamed Gorgon Halo and will be a similar affair as the Gorgon Point refresh of mainstream APUs. It should therefore use the same 4nm silicon as Strix Halo, meaning there will be no changes to the architectures of the CPU cores, GPU core or other blocks, and core counts also remain fixed. As a result, what the refresh can bring is primarily higher clock speeds.<\/p>\n Following the VideoCardz report, the arrival of these refresh models was also confirmed by the Hong Kong website HKEPC, which also added preliminary specifications of the upcoming models, including their processor codes. These indicate that Gorgon Halo will essentially mirror the Strix Halo model lineup, with two eight-core models, two twelve-core models, and one sixteen-core model. The sixteen-core variant will feature a fully enabled Radeon 8060S GPU with 2560 shaders, while the twelve-core and eight-core models will be available either with the Radeon 8060S or with a cut-down Radeon 8050S GPU, which has \u201conly\u201d 2048 shaders.<\/p>\n Even the model numbers will be the same, just starting with four at the beginning instead of a three. According to HKEPC, this will effectively be a rebranding for most of the SKUs, as clock speeds do not differ at all for the four lower-end models\u2014the Radeon 8060S runs at 2.9 GHz and the Radeon 8050S at 2.8 GHz. The maximum boost clocks of the Zen 5 cores are also said to be unchanged\u2014officially given as 5.0 GHz. The processors could, however, reportedly have better efficiency and thus achieve higher multithreaded performance at the same power consumption.<\/p>\n The only model where the refresh will have any visible impact on specifications is the top-tier 16-core Ryzen AI Max+ 495. Compared to the currently sold 395 version, it has a higher listed boost\u2014the maximum turbo will be 5.2 GHz instead of 5.1 GHz. This will allow the processor to achieve a few (literally just a few) percent higher scores in single-threaded benchmarks, but the resulting difference won’t be particularly decisive in practice. The GPU clock will also increase slightly\u2014for this specific processor, the integrated GPU will run at 3.0 GHz.<\/p>\n
Model<\/strong><\/span><\/td>\n| Cores<\/strong><\/span><\/td>\n | L3<\/strong><\/span><\/td>\n | Clock speed<\/strong><\/span><\/td>\n | Turbo<\/strong><\/span><\/td>\n | GPU: shaders<\/strong><\/span><\/td>\n | GPU clock speed<\/strong><\/span><\/td>\n | TDP<\/strong><\/span><\/td>\n | OPN<\/strong><\/span><\/td>\n<\/tr>\n<\/tbody>\n\n | Ryzen AI MAX+ 495<\/td>\n | 16 (32 v.)<\/td>\n | 2\u00d732 MB<\/td>\n | 3,1 GHz<\/td>\n | 5,2 GHz<\/td>\n | Radeon 8060S<\/td>\n | 3,0 GHz<\/td>\n | 55 W<\/td>\n | 100-000002145<\/td>\n<\/tr>\n | Ryzen AI MAX 492<\/td>\n | 12 (24 v.)<\/td>\n | 2\u00d732 MB<\/td>\n | 3,2 GHz<\/td>\n | 5,0 GHz<\/td>\n | Radeon 8060S<\/td>\n | 2,9 GHz<\/td>\n | 55 W<\/td>\n | 100-000002143<\/td>\n<\/tr>\n | Ryzen AI MAX 490<\/td>\n | 12 (24 v.)<\/td>\n | 2\u00d732 MB<\/td>\n | 3,2 GHz<\/td>\n | 5,0 GHz<\/td>\n | Radeon 8050S<\/td>\n | 2,8 GHz<\/td>\n | 55 W<\/td>\n | 100-000002142<\/td>\n<\/tr>\n | Ryzen AI MAX 488<\/td>\n | 8 (16 v.)<\/td>\n | 36 MB<\/td>\n | 3,2 GHz<\/td>\n | 5,0 GHz<\/td>\n | Radeon 8060S<\/td>\n | 2,9 GHz<\/td>\n | 55 W<\/td>\n | 100-000002140<\/td>\n<\/tr>\n | Ryzen AI MAX 485<\/td>\n | 8 (16 v.)<\/td>\n | 36 MB<\/td>\n | 3,2 GHz<\/td>\n | 5,0 GHz<\/td>\n | Radeon 8050S<\/td>\n | 2,8 GHz<\/td>\n | 55 W<\/td>\n | 100-000002127<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n | AMD will therefore use this refresh mainly to slightly speed up the flagship model, which may be achieved by selecting the highest-quality silicon for this single SKU. The other models look more like a simple rebranding\u2014unless there’s some surprise hiding in other unpublished parameters. The Gorgon Halo models could support LPDDR5X-8533 memory in theory, which would improve GPU bandwidth somewhat\u2014the 300 generation supports only LPDDR5X-8000 speed.<\/p>\n Medusa Halo with Zen 6 and RDNA 5 not until two years from now<\/h2>\nThis refresh \u201chalf-generation\u201d will arrive on the market later than the mainstream Ryzen AI 400 \u201cGorgon Point\u201d. It is expected to appear only in the last quarter of this year. It will likely remain on the market throughout the entirety of 2027 after that.<\/p>\n According to information from the leaker known as Kepler_L2, the next completely new generation in the \u201cHalo\u201d processor lineup will arrive quite late. Medusa Halo processors, which will feature Zen 6 cores and integrated graphics based on the RDNA 5 generation, are reportedly planned for release only during CES 2028, that is, in January two full years from now. Gorgon Halo will therefore have to serve as the high-end laptop processor until that time.<\/p>\n Sources: VideoCardz (1<\/a>, 2<\/a>), HKEPC<\/a>, <\/em>Kepler_L2<\/em><\/a><\/p>\n English translation and edit by Jozef Dud\u00e1\u0161<\/em><\/p>\n |