MiniMax officially launched M3.1 Flash on October 19, 2026, and confirmed that the anonymous “Space Bunny” model that topped OpenRouter in stealth was an early version of it. Why it matters: M3.1 Flash pairs frontier-style agentic and coding claims with very low API pricing — $0.15 per million input tokens and $0.60 per million output tokens.
October 19, 2026
What happened
MiniMax describes M3.1 Flash as a next-generation Flash model with a generational leap in agentic tasks, coding, computer use, and cybersecurity. The company says online inference is nearly three times faster than its predecessor, while usage costs for agentic tasks are 70% lower.
OpenRouter, which hosted the stealth run, said Space Bunny was its number-one model for 10 straight days and processed 56.7 trillion tokens, the most of any stealth model it has run. OpenRouter said feedback from that period helped MiniMax cut time to first token by 38%, raise throughput by 45%, and halve the error rate. It is offering M3.1 Flash free for 72 hours.
What MiniMax claims on benchmarks
These figures are company-reported and have not been independently verified. MiniMax reported 73.7% on DeepSWE v1.1 for complex software work, 76.2 on BankerToolBench for multi-step financial workflows, 61.7 points on OSWorld V2 for computer use, and 82% on CyberGym for vulnerability discovery and validation.
The company also said the model can sustain output speeds above 170 tokens per second at the median in everyday use, with speeds exceeding 200 tokens per second in some conditions, helped by more aggressive sparsity.
Pricing and availability
Flat API pricing is $0.15 per million input tokens, $0.60 per million output tokens, and $0.015 per million cached input tokens. The model is available through the MiniMax API, the company’s M Plan subscription, and MiniMax Code for desktop, as well as through OpenRouter.
MiniMax’s launch post said the model’s weights will arrive on October 23, 2026. The company also said a Pro model will arrive later.
Why it matters
Cheap, fast models are what make always-on agents affordable. If M3.1 Flash holds up outside MiniMax’s own tests, its price puts pressure on every other lab selling high-volume agent and coding work. The honest test is the same as always: run it on your own tasks, not the launch benchmarks. The weights release on October 23 will also tell us how open this launch really is.
Frequently asked questions
What is MiniMax M3.1 Flash?
It is MiniMax’s new Flash model for agentic tasks, coding, computer use, and cybersecurity, launched on October 19, 2026.
Was Space Bunny really M3.1 Flash?
Yes. MiniMax confirmed Space Bunny, the anonymous model that ran in stealth on OpenRouter, was an early version of M3.1 Flash.
How much does M3.1 Flash cost?
MiniMax lists $0.15 per million input tokens, $0.60 per million output tokens, and $0.015 per million cached input tokens. OpenRouter offered it free for 72 hours at launch.
Sources: Unite.AI reporting on MiniMax’s announcement and OpenRouter’s reveal (October 19, 2026)