MiniMax M3.1 Flash
A lightweight, high-speed text model released by MiniMax, targeting everyday development scenarios. It is designed for low latency and high response speed, making it suitable for high-frequency tasks such as rapid code fixes and similar lightweight workloads.
Core Facts
M3.1 Flash can be called via the AI SDK, Chat Completions, Responses, and the Anthropic Messages API, with the model name minimax/minimax-m3.1-flash.
Maximum output tokens: MiniMax M3.1 Flash is 512K, DeepSeek V4.1 Flash 1017 is 384K, and Claude Haiku 5.5 is 128K.
MiniMax M3.1 Flash offers a 1M-token ultra-long context window.
Timeline (last 90 days)
ClinePass added Kimi K3.1 Pro (which became the default), MiniMax M3.1 Flash, and DeepSeek V4.1 Flash 1017, and removed Mimo V2.5 and V2.5 Pro.
On the uploader's RTX Pro 5000 platform, the speed performance of each model on Xtier is: Qwen 3.8 is the fastest with essentially no waste; DeepSeek 5.3 is slightly slower by 2% than the revised LS济浪; GLM 5.3 is faster than LS济浪; MiniMax M3.1 Flash is very slow.
The active expert data volume of GLM 5.3 is 8.8GB, smaller than the 13.2GB of MiniMax M3.1 Flash.
MiniMax M3.1 Flash launched, combining with SGLang Omni to provide the foundation for building voice agents; the pipeline uses a local Qwen 3 ASR for speech recognition and a local Qwen 3 TTS for speech synthesis, with the LLM stage calling MiniMax API's M3.1 Flash; integration testing ran on 1xH200, with interrupt timing at 0.17 seconds, average stop latency at 1.32 seconds, and 16 concurrent drawings on a single GPU.
The MiniMax M3.1 Flash model has the internal codename "Space Bunny".
The model's actual performance still needs community validation, especially its accuracy and safety in coding and cybersecurity tasks.
MiniMax M3.1 Flash targets scenarios such as agentic tasks, coding, computer use, and cybersecurity.
Together AI announced that the MiniMax M3.1 Flash model is now officially available on its platform, launched simultaneously as a Day-0 launch partner.
Between the two backup models, M3.1 was preferred over M3.
The current configuration of "Taka Story" is DeepSeek as primary and M3.1 as backup.
40 more timeline events
All Facts (20)
M3.1 Flash can be called via the AI SDK, Chat Completions, Responses, and the Anthropic Messages API, with the model name minimax/minimax-m3.1-flash.
75%VerifiedMaximum output tokens: MiniMax M3.1 Flash is 512K, DeepSeek V4.1 Flash 1017 is 384K, and Claude Haiku 5.5 is 128K.
70%VerifiedMiniMax M3.1 Flash offers a 1M-token ultra-long context window.
65%VerifiedThe official name of MiniMax M3.1 Flash is `MiniMax-M3.1-Flash`, positioned as a frontier multimodal coding model with a 1 million context window and adjustable thinking depth.
65%VerifiedMiniMax M3.1 Flash Preview emphasizes speed and can be used for quickly fixing bugs in code development.
65%Partially VerifiedClinePass added Kimi K3.1 Pro (which became the default), MiniMax M3.1 Flash, and DeepSeek V4.1 Flash 1017, and removed Mimo V2.5 and V2.5 Pro.
80%Partially VerifiedMiniMax M3.1 Flash is available on AIHubMix for token-billed calls under the model ID `minimax-m3.1-flash`, with input priced at $0.165 per million tokens and output at $0.66 per million tokens.
65%Partially VerifiedCalculated per thousand repair tasks, Claude Haiku 5.5 costs about $0.70, DeepSeek V4.1 Flash 1017 about $1.35, and MiniMax M3.1 Flash about $4.45 by default and about $0.64 at the low setting.
65%UnverifiedMiniMax M3.1 Flash uses a MoE architecture with approximately 428B total parameters and approximately 23B activated per token.
60%UnverifiedIn the dual-bug Python function repair benchmark (215 hidden tests), omitting effort (max) consumes an average of 6,364 reasoning tokens, takes 56 seconds, and costs $0.0045 per run; effort=low consumes an average of 591 tokens, takes 9.5 seconds, and costs $0.0006 per run.
60%UnverifiedMiniMax M3.1 Flash defaults to the max setting when the effort field is not passed in the request, and the model always performs thinking.
60%UnverifiedWhen the reasoning_effort field is not passed, MiniMax M3.1 Flash runs at the max thinking level, taking an average of 56.3 seconds for a small repair task, which may trigger client timeouts tuned for fast models.
60%UnverifiedThe active expert data volume of GLM 5.3 is 8.8GB, smaller than the 13.2GB of MiniMax M3.1 Flash.
50%UnverifiedOn the uploader's RTX Pro 5000 platform, the speed performance of each model on Xtier is: Qwen 3.8 is the fastest with essentially no waste; DeepSeek 5.3 is slightly slower by 2% than the revised LS济浪; GLM 5.3 is faster than LS济浪; MiniMax M3.1 Flash is very slow.
50%UnverifiedMiniMax M3.1 Flash launched, combining with SGLang Omni to provide the foundation for building voice agents; the pipeline uses a local Qwen 3 ASR for speech recognition and a local Qwen 3 TTS for speech synthesis, with the LLM stage calling MiniMax API's M3.1 Flash; integration testing ran on 1xH200, with interrupt timing at 0.17 seconds, average stop latency at 1.32 seconds, and 16 concurrent drawings on a single GPU.
50%UnverifiedThe MiniMax M3.1 Flash model has the internal codename "Space Bunny".
50%UnverifiedThe model's actual performance still needs community validation, especially its accuracy and safety in coding and cybersecurity tasks.
50%UnverifiedMiniMax M3.1 Flash targets scenarios such as agentic tasks, coding, computer use, and cybersecurity.
50%UnverifiedThe current configuration of "Taka Story" is DeepSeek as primary and M3.1 as backup.
50%UnverifiedAfter optimization, the new round of M3.1 had an average response time of 16.11 seconds, but the long-tail problem remains unresolved.
50%