MiniMax M3 Is Now Available in Puter.js

Puter.js now supports MiniMax M3, the latest frontier-level model from MiniMax — the first open-weight model to combine frontier coding, a 1-million-token context window, and native multimodality in a single model.

What is MiniMax M3?

MiniMax M3 is a multimodal language model released on June 1, 2026, built for long-horizon coding, agentic workflows, and complex reasoning. Its headline innovation is MiniMax Sparse Attention (MSA) — an architecture that selects only the relevant blocks of the KV cache for each token, slashing per-token compute at long context to roughly 1/20 of the prior generation. Key highlights include:

• Frontier Coding: Scores 59.0% on SWE-Bench Pro, surpassing GPT-5.5 and Gemini 3.1 Pro and approaching Claude Opus 4.7
• Best-in-Class Agentic Tasks: Highest score on Claw-Eval (74.5%), with 66.0% on Terminal-Bench 2.1 and 74.2% on MCP Atlas
• 1M-Token Context with MSA: Up to 9× faster prefill and 15× faster decoding at 1M tokens compared to the previous generation
• Native Multimodality: Accepts text, image, and video inputs natively from mixed-modality training
• Standard Tool Calling: Works with standard MCP scaffolding for coding agents and multi-step automation

Examples

Long-horizon coding task

puter.ai.chat("Build a CLI tool in Go that watches a directory, debounces file changes, and runs a configurable build command. Include graceful shutdown, structured logging, and unit tests.",
  { model: 'minimax/minimax-m3' }
);

Agentic workflow

puter.ai.chat(`Plan and implement an autonomous agent that:
1. Monitors a GitHub repo for failing CI runs
2. Pulls the relevant logs and diffs
3. Proposes a fix with tests
4. Opens a pull request with a clear summary`,
  { model: 'minimax/minimax-m3', stream: true }
);

Long-document analysis

puter.ai.chat("Summarize this 400-page technical specification, extract every API endpoint with its parameters, and flag any inconsistencies between sections.",
  { model: 'minimax/minimax-m3', stream: true }
);

Multimodal input

puter.ai.chat(
  "Look at this UI screenshot and generate the React component that reproduces it, including responsive styling.",
  "https://assets.puter.site/doge.jpeg",
  { model: 'minimax/minimax-m3' }
);

Get Started Now

Just add one library to your project:

// npm install @heyputer/puter.js
import { puter } from '@heyputer/puter.js';

Or add one script tag to your HTML:

<script src="https://js.puter.com/v2/"></script>

No API keys needed. Start building with MiniMax M3 immediately.

Learn more:

• Free, Unlimited MiniMax API
• Puter.js AI Documentation