MiniMax officially launches M3.1 Flash, with a minimum price of $0.15 per million tokens, and it will be open-sourced on the 23rd.

Beating AI Express: MiniMax officially launches M3.1 Flash, confirming that the previously viral anonymous model Space Bunny is its early version. Space Bunny launched on OpenRouter and OpenCode in late September, with total calls exceeding 200 trillion tokens and once topping the daily call charts on both platforms. MiniMax rolled out M3.1 Flash Preview on September 27, and has now officially released full benchmark results and API pricing. In MiniMax’s published tests, M3.1 Flash scored 62.2% on the office automation benchmark Automation-Bench, outperforming DeepSeek V4.1 Flash (60.5%) and Opus 5 (50.3%). It notched 73.7% on the DeepSWE programming test, close to Opus 5’s 74.0%; its visual reasoning score on BabyVision hit 91.0%. However, its 61.7% on the computer operation test OSWorld still lags behind Opus 5’s 70.6%. Pricing is a key highlight: the API charges $0.15 per million input tokens, $0.60 per million output tokens, and just $0.015 for cached reads. MiniMax states that compared to M3, the new model’s overall cost is reduced to roughly one-third, while its speed is tripled. The model supports a million-token context window, with weights scheduled to open on October 23.