MiniMax-M3
No published training cutoff
MiniMax-M3
$0.30 in, $1.20 out per 1M tokens
Successor to the plan-only MiniMax-M3.1-Flash-Preview (Sept 27). Same Anthropic-format endpoint as MiniMax-M3.
Model idMiniMax-M3.1-Flash
Price$0.15 in, $0.60 out per 1M tokens
Checked19 October 2026
At least0events since its release on 19 October 2026
0 historic, 0 major. 0 days of news.
MiniMax doesn’t publish a training cutoff, so we count from its release date. The real gap is larger.
3 ways to use MiniMax-M3.1-Flash, with the model id each provider expects.
| Provider | Model id | Endpoint | Docs |
|---|---|---|---|
| MiniMax API | MiniMax-M3.1-Flash |
https://api.minimax.io/anthropic |
platform.minimax.io |
| OpenRouter | openrouter.ai | ||
| MiniMax Code (desktop) | agent.minimax.io |
| Input | $0.15 per 1M tokens |
|---|---|
| Output | $0.60 per 1M tokens |
| Cache read | $0.015 per 1M tokens |
As published: “per 1M tokens (USD), flat rate”.
Source: minimax.io, checked 19 October 2026.
What its maker marketed at launch, and what people found later, each with its source.
$0.15/$0.60 per 1M tokens, 200+ tokens/s; MiniMax reports DeepSWE v1.1 73.7%, OSWorld V2 61.7, Agents’ Last Exam 29.6, CyberGym 82% (self-reported).
Ran anonymously on OpenRouter as Space Bunny, #1 by usage for 10 days with 56.7T tokens, before the reveal.
Weights promised for Oct 23, 2026 (MiniMax X post); until then API-only, so open_weights is false and license unknown. Context window and max output not stated at launch. Image input inferred from MiniMax’s BabyVision score; not confirmed in docs.
Sources: https://www.minimax.io/blog/minimax-m31f · https://x.com/MiniMax_AI/status/2112150867868541221
Not yet checked against an official source: OpenRouter slug not verified.
Paste this into a new chat. It points your AI to a dated list of AI news since September 2026, with a source for every item.
Using Gemini, Grok or another assistant? Copy the prompt and paste it into a new chat. Only Claude and ChatGPT have a link that opens with a prompt filled in.
Files for every cutoff, with their sizes: Briefings.
For developers: an MCP server is planned. For AI agents