Key Info

MiniMax announced a strategic partnership with Inferact to advance production inference for MiniMax models, with engineering improvements contributed back to the vLLM project. Inferact has served MiniMax M3.1 Flash in production since its launch day.

Highlights

  • Inferact has run MiniMax M3.1 Flash in production since day one of launch.
  • Requests through the MiniMax platform are routed to Inferact's stack.
  • Engineering improvements from the collaboration are contributed back to vLLM.
  • The partnership targets frontier production inference for open-source models.