MiniMax Launches M3.1 Flash, Cutting Agent Task Costs by 70% Compared to Predecessor

MiniMax has officially released its new-generation model M3.1 Flash. The usage cost for agent tasks is reduced by 70% compared to the previous generation, online inference speed is nearly tripled, and the output speed exceeds 200 tokens per second. The API is priced at $0.15 per million input tokens and $0.6 per million output tokens. This model scores 61.7 on the OSWorld V2 benchmark, ranking first among domestic Chinese models. The official confirmed that the anonymous model "Space Bunny" which previously processed over 200 trillion tokens is an early version of M3.1 Flash. The model is now available via MiniMax API, M Plan and MiniMax Code, with model weights to be open-sourced at a later time.

Source: PANewsRead original