MiniMax's M3.1 Flash is now live on OpenRouter at $0 prompt / $0 completion for a limited-time launch window - free for the next 72 hours from October 19. OpenRouter also revealed the model is "Space Bunny", the stealth model that topped OpenRouter's daily volume charts for 10 straight days, processing 56.7T tokens - the most of any stealth model they've run.
Verified specs (from OpenRouter's live API)
- Model id:
minimax/minimax-m3.1-flash:free(canonical:minimax-m3.1-flash-20261019) - $0 per token on prompt and completion during the free window
- 1M context window, up to 524K max output
- Multimodal input: text, image, and video
- Mandatory reasoning with efforts from low to max (default: max)
- Tool calling, structured output, response format supported
- Model went live October 19, 2026
What it is
MiniMax-M3.1-Flash is MiniMax's fast multimodal frontier model, built for complex multi-step agentic work, coding, and cybersecurity - the GA version of the M3.1-Flash-Preview that had been in testing since late September.
How to use it
- Get an OpenRouter key (openrouter.ai)
- Point your OpenAI-compatible client at
https://openrouter.ai/api/v1 - Set model to
minimax/minimax-m3.1-flash:free
Caveats
72-hour window: free access started October 19; after it ends, the paid endpoint applies (the campaign's listed price is $0.15/M in, $0.60/M out)
Free endpoints on OpenRouter are rate-limited (see their rate limit docs)
Errors and empty responses still count against limits
Free endpoint: openrouter.ai/minimax/minimax-m3.1-flash:free