[MiniMax Releases M3.1 Flash, Reducing Costs to One-Third of M3 and Tripling Speed] According to Deep Tide TechFlow, on October 19, MiniMax announced the official release of M3.1 Flash. The early version of this model, Space Bunny, which previously topped global rankings and processed over 200T tokens cumulatively, served as its predecessor. M3.1 Flash achieved scores of 62% on Automation-Bench, 30% on ALE, 74% on DeepSWE, and 91% on BabyVision for tasks such as Cowork, Coding, and Visual Reasoning. MiniMax stated that the cost of M3.1 Flash is approximately one-third of M3, with speeds tripled. The API is uniformly priced at $0.15/$0.60 per 1 million input/output tokens, and cache read costs are $0.015 per 1 million tokens. Model weights will be released on October 23.