MiniMax releases M3.1 Flash at $0.15 per million input tokens
October 20: Latent Space says MiniMax expects M3.1 Flash weights on October 23. October 19 reports add the 428B MoE description and detailed benchmark and API prices.
MiniMax has released M3.1 Flash, a cheap reasoning model scoring 42 on Artificial Analysis, with weights said to arrive October 23.
MiniMax released M3.1 Flash, a reasoning model MarkTechPost describes as a 428-billion-parameter mixture-of-experts agentic model. The Decoder reports an API price of $0.15 per million input tokens, $0.60 per million output tokens, and $0.015 for cached input, with about $0.22 per task and a one-million-token text-and-image context. MiniMax says it scores 42 on the Artificial Analysis Intelligence Index, 13 points above M3 and less than half the cost per task, tying GLM-5.3-Flash and trailing Deepseek V4.1 Flash 1017 at 45. Artificial Analysis reports a 93 percent hallucination rate because the model answers every question. Latent Space says MiniMax expects weights on October 23; the API is already live.
- Input $0.15, output $0.60, cached input $0.015 per million tokens.
- Artificial Analysis Intelligence Index score of 42, up 13 from M3.
- Humanity's Last Exam 48 percent; CritPt physics 22 percent on 71 problems.
- DeepSWE v1.1 up to 73.7 percent at under $0.30 per task.
- MiniMax says the model is three times faster than M3.
- MarkTechPost calls it a 428B MoE; retrieved page lacks specs.
Timeline
dated events in the reporting · oldest first- Oct 23, 2026MiniMax says M3.1 Flash weights are due · Latent Space
What each outlet adds
beyond the shared facts- The DecoderIndex, exam scores, hallucination rate, context length, and API prices.
- MarkTechPostDescribes M3.1 Flash as a 428B MoE; page lacks specs.
- Latent SpaceMiniMax says open weights arrive October 23.
- Products
- MiniMax Code
- Organizations
- Artificial AnalysisTogether
- Countries
- China
Coverage
oldest first · one row per article · UTC- · 6d ago
MarkTechPostFirst report
MiniMax launches M3.1 Flash, a 428B MoE agentic model
The listing prices input at $0.15 per million tokens; the fetched page body is only a cookie notice.
- · 6d ago
The Decoder
MiniMax releases M3.1 Flash, with open weights due October 23
API pricing is $0.15 per million input tokens and $0.60 per million output tokens, with cached input at $0.015.
- · 5d ago
Latent Space
DeepSeek V4.1 Flash 1017 leads open-weight value rankings
Artificial Analysis scores the 552B MoE at 45, up six points, tying GLM-5.3 among open weights.