MiniMax M3.1 Flash officially launches

MiniMax officially updated M3.1 Flash on October 19; Space Bunny, which anonymously processed more than 200 trillion tokens, was an early version. The company says the model delivers a generational improvement across complex agent tasks, coding, computer use and cybersecurity, with online inference nearly three times faster and total usage cost on agentic tasks 70% lower than the previous generation. Published scores include 76.2 on BankerToolBench, 73.7% on DeepSWE v1.1, 82% on CyberGym and 61.7 on OSWorld V2. It supports a one-million-token context plus text, image and video input through MiniMax Code, M Plan and the developer platform. Speed, cost and benchmarks are vendor claims; tool success, stability, safety boundaries and total cost still need independent testing on real workflows.