On January 25th, AMD announced that it has integrated a new DeepSeeker V3 model onto the Instinct MI300X GPU, which is designed to achieve optimal performance together with SGLang. DeepSeeker V3 has been optimized for Al inference.
Scan code to share