MiniMax M3.1 Flash Gets Day-One Support from SGLang and Miles
MiniMax announced that M3.1 Flash is ready from inference to RL, with SGLang and Miles providing support on the very first day of release. According to SGLang, the model achieves over 1K tok/s for bs1 decode on SGLang, supports NVIDIA Blackwell and Hopper (GB300/B300/B200/H200) as well as AMD Instinct MI350X and MI355X, and supports MiniMax's DSpark draft model along with HiCache and multimodal input.
Why read this
AI score 55
The original text provides M3.1 Flash's throughput, hardware coverage, and speculative decoding support on SGLang, which can be used to assess deployment costs and compatibility scope.
Topics
View original postSpace Bunny is here! 🐰 MiniMax M3.1 Flash hits >1K tok/s, bs1 decode with SGLang, production-ready from Day 0. This integration is a joint effort by @MiniMax_AI, @NVIDIAAI, @AMD, and @radixark, and you can also run RL with Miles! -Broad hardware support: NVIDIA Blackwell and Hopper (GB300/B300/B200/H200), plus AMD Instinct MI350X and MI355X. -Agentic serving at scale: ~100K total tok/s per GPU on the AgentX coding-agent trace with 96 concurrent sessions. -Speculative decoding: Supports MiniMax's DSpark draft and a DFlash2 draft trained by @radixark; supports HiCache and multimodal input. -Ultra-low-latency inference: MegaKernels deliver 2.15× faster single-request decode.
View original
Space Bunny is here! 🐰 MiniMax M3.1 Flash hits >1K tok/s, bs1 decode with SGLang, production-ready from Day 0. This integration is a joint effort by @MiniMax_AI, @NVIDIAAI, @AMD, and @radixark, and you can also run RL with Miles! -Broad hardware support: NVIDIA Blackwell and Hopper (GB300/B300/B200/H200), plus AMD Instinct MI350X and MI355X. -Agentic serving at scale: ~100K total tok/s per GPU on the AgentX coding-agent trace with 96 concurrent sessions. -Speculative decoding: Supports MiniMax's DSpark draft and a DFlash2 draft trained by @radixark; supports HiCache and multimodal input. -Ultra-low-latency inference: MegaKernels deliver 2.15× faster single-request decode.
Body
From inference to RL, ready from day one. 🐰
Thanks to the teams behind SGLang, and Miles @sgl_project alongside @NVIDIAAI and @AMD, for supporting MiniMax M3.1 Flash. Great building together!