MiniMax-M3-MXFP8

Papers using or citing this model 1

• MiniMax Sparse Attention 36 citations · 2026 · arXiv (Cornell University) · arXiv

  Ultra-long-context capability is becoming indispensable for frontier LLMs: agentic workflows, repository-scale code reasoning, and persistent memory all require the model to jointly attend over hundreds of thousands to millions of...