MiniMax M3
MiniMax M3 is a multimodal foundation model developed by MiniMax that accepts text, image, and video inputs. It features a large 1 million-token context window and supports reasoning and tool-calling capabilities, making it suitable for complex, long-horizon tasks.
MiniMax M3 specs and pricing
- Provider: MiniMax
- Released: May 31, 2026
- Context window: 1M tokens
- Max output: 512K tokens
- Input: text, images, video
- Output: text
- Price: $0.30 input / $1.20 output per 1M tokens
- Reasoning: yes
- Tool calling: yes
What MiniMax M3 is good at
- Analyzing long documents or complex codebases
- Processing multimodal data including images and video
- Executing multi-step agentic workflows
To-dos to try with MiniMax M3
- Upload a long video of a technical presentation and summarize the key points.
- Paste a large codebase and ask the model to identify potential bugs or architectural improvements.
- Provide a series of images and a text prompt to generate a detailed analysis of the visual content.
FAQ
- What types of input does MiniMax M3 support? The model supports text, image, and video inputs.
- What are the primary capabilities of this model? MiniMax M3 is designed for reasoning and tool-calling, which allows it to perform complex, multi-step tasks.
- How much does MiniMax M3 cost? MiniMax M3 costs $0.30 per million input tokens and $1.20 per million output tokens through OpenRouter. In todo.is a typical agent step costs about 1.01 credits ($1 = 100 credits).
- What is the context window of MiniMax M3? MiniMax M3 can read up to 1M tokens at once and write up to 512K tokens in one reply.
- Can I use MiniMax M3 without coding? Yes. In todo.is you pick MiniMax M3 as your agent's model and write to-dos in plain words, on the web, the Android app, WhatsApp or Telegram. No API key needed: credits are included, or you can bring your own key.
JavaScript is required to use the todo.is app.