vLLM, AMD and Embedded LLM optimize MiniMax M3 on AMD Instinct MI355X
Published performance-engineering results from optimizing MiniMax M3 on the AMD Instinct MI355X. Company-reported: speedups of 3.14 times at MXFP8 concurrency 32 (342.4 tok/s per GPU), 2.09 times at concurrency 128 (623.7 tok/s) and 4.45 times with MXFP4 (TP2/EP1) (943.5 tok/s), and a total of 6,370.5 tok/s/GPU in a prefill/decode disaggregated setup. Fusing the shared expert into the routed-expert path cut kernel launches in the low-concurrency range by 30.2%; a compact page table that fits 128-token block sparse attention onto 16-token pages eliminated KV-cache copies; and extending AITER sparse attention to multi-token validation improved speculative decoding throughput by 7.9–8.3%.
Why it matters
A record, in concrete numbers, of AMD GPUs being optimized in a major open inference engine into a real alternative to NVIDIA.
Source — Primary sources first; secondary reports are labelled “Press”
Officialvllm.aihttps://vllm.ai/blog/2026-09-10-minimax-m3-mi355xAdded in Week 38, 2026 · See that week