Breaking01 / AI
MiniMax unveils M3.1 Flash at $0.15 per million input tokens
MiniMax claims a 70% cost cut for agentic tasks but has not disclosed its previous model's prices or when the weights will be released.

MiniMax released M3.1 Flash on 19 Oct 2026 at $0.15 per million input tokens and $0.60 per million output tokens for API users, the company announced.
Pricing. MiniMax said the overall usage cost of agentic tasks is 70% lower than on its previous model. The price table in the announcement lists $0.015 per million cached input tokens, and pricing is flat.
The announcement confirms that the model is available through the MiniMax API, the M Plan subscription and MiniMax Code. Output speed reaches 200-plus tokens per second, with a median of 170-plus tokens per second, according to company-published figures.
Company-published figures in the announcement put M3.1 Flash at 73.7% on DeepSWE v1.1, a software development benchmark. DeepSWE measures whether a model can understand code, develop features, fix issues and validate its work through testing on its own. The company-published score on CyberGym, a cybersecurity benchmark, is 82%.
The company-published score on OSWorld V2, a computer-use benchmark, is 61.7 points. On Agents’ Last Exam, a benchmark of professional work, the company-published score is 29.6 points. MiniMax said the model ranks first among Chinese models on both.
MiniMax wrote in the announcement that the Flash line is an effort to balance efficiency and capability. It said Space Bunny, a model that processed more than 200 trillion tokens while operating anonymously, was an early version of M3.1 Flash.
The company said it would release model weights and a Pro model later. MiniMax has not disclosed the release date for the weights or the prices of its previous model.