AMD has officially introduced the Ryzen AI MAX 400 series of system-on-chips. The lineup belongs to the Gorgon Halo family and targets high-end artificial intelligence workloads. AMD released three distinct models within this new generation: the Ryzen AI MAX+ PRO 495, the Ryzen AI MAX+ PRO 490, and the Ryzen AI MAX+ PRO 485.
New SoCs enable local execution of massive language models with high memory capacity.
The architecture supports up to 192 GB of unified memory. This large memory capacity allows a single chip to run local language models with over 300 billion parameters. The flagship Ryzen AI MAX+ PRO 495 integrates sixteen Zen 5 CPU cores and a Radeon 8065S integrated graphics processor featuring forty compute units. AMD increased the maximum boost clock speed by 100 MHz compared to the previous generation, reaching up to 5.2 GHz. The dedicated neural processing unit delivers fifty-five tera-operations per second of AI performance.
Manufacturers including ASUS, HP, and Lenovo will release devices powered by these chips in the third quarter of 2026.
The Gorgon Halo series positions AMD directly within the competitive landscape for local AI inference hardware. The combination of high memory bandwidth and significant neural processing power addresses a growing demand for on-device large language model execution. This architecture enables enterprise and enthusiast systems to process massive models without relying entirely on cloud infrastructure.
Discussion
0 comments
Log in to join the thread with a thoughtful take, question, or correction.