Unverified50% confidenceFactExact time
Early users report that MiniMax M3.1 Flash decodes at 90–110 tokens/second, but MiniMax has not yet published official speed data.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
10/19/2026
First Seen
Valid until: 11/2/2026
Sources
Related Entities
Related Claims
UnverifiedMiniMax M3.1 Flash's thinking tokens count toward the max_tokens limit, so reusing the 4K or 8K caps from the M2.x era may cause truncation with finish_reason: "length"; MiniMax recommends raising max_tokens to 131,072, with an upper limit of 524,288.75% similarUnverifiedMaximum output tokens: MiniMax M3.1 Flash is 512K, DeepSeek V4.1 Flash 1017 is 384K, and Claude Haiku 5.5 is 128K.75% similarUnverifiedMiniMax M3.1 Flash offers a 1M-token ultra-long context window.74% similarUnverifiedMiniMax M3.1 Flash defaults to the max setting when the effort field is not passed in the request, and the model always performs thinking.73% similarVerifiedDeveloper Primitive reported a generation speed of 142.6 tokens/s for Qwen3 Flash Next running locally.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1068556API
curl https://kongchang.com/api/v1/knowledge/claims/1068556MCP
get_claim(id=1068556)