Unverified50% confidenceReleaseExact time
MiniMax M3.1 focuses on optimizing stability, reasoning quality, efficiency, and agent generalization, while M3 Pro takes a scale-up approach and reportedly has nearly 3 trillion parameters.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/23/2026
First Seen
Valid until: 12/22/2026
Sources
Related Entities
Related Claims
UnverifiedThe model used in the demonstration is MiniMax M3, not the highest-performing model on the market; the selection criteria were sufficiency and cost reliability.79% similarUnverifiedMiniMax M2.7采用230亿参数的稀疏MoE（混合专家）架构，每个token仅激活100亿参数74% similarUnverifiedOptimizing MiniMax M3 inference serving on the MI355X ultimately achieved a 3–4x improvement in serving throughput.73% similarUnverifiedMiniMax M3 scored 60 points in the drag-and-drop kanban test, with complete drag-and-drop logic but lacking internationalization support and responsive design.68% similarUnverifiedThe ratio of activated parameters to total parameters for MiniMax M3, 23B/428B, is approximately 5.4%.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/942576API
curl https://kongchang.com/api/v1/knowledge/claims/942576MCP
get_claim(id=942576)