MiniMax-M3
MiniMax's open-weight flagship, built for long-context and agentic workloads with a very large context window and strong tool use at a low served price.
At a glance
Released
Jun 3, 2026
Provider
MiniMax
Context window
1,000,000 tokens
Max output
32,000 tokens
Modalities
Text + image → text
Knowledge cutoff
February 2026
API model ID
minimax-m3License
Open weights (MiniMax Model License)
Pricing
Standard
Input$0.29 / 1M
Output$1.16 / 1M
Long-context requests are billed at $0.58 in / $2.32 out per 1M. A typical 1,500-token prompt with a 650-token answer costs ≈ $0.001 on the standard tier. Run your own prompt through the calculator →
What's new vs the previous model
Compared withMiniMax-M2 (December 2025)
- Much longer context and stronger agentic tool use
- Better coding and reasoning than M2
- Higher serving cost than the previous generation
Best use cases
Long-context agents
Sustains long-horizon agent sessions with large working memory.
Document-heavy RAG
Reasons over very large corpora in a single context window.
Open deployment
Self-hosts for teams needing frontier-style long context on their own stack.
Strengths & limitations
Strengths
- Huge context
- Strong agentic tool use
- Open weights
Limitations
- Resource-intensive at full context
- Ecosystem smaller than Llama
How to test it
- No-code: open chat.minimax.io and pick MiniMax-M3 from the model switcher — no API key needed.
- Playground: sign in to the MiniMax Open Platform and run prompts in the workbench, then copy the generated code.
- API: create a key in the MiniMax Open Platform and call the API (example below).
curl https://api.minimax.io/v1/chat/completions \
# OpenAI-compatible endpoint
-H "Authorization: Bearer $API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "minimax-m3",
"messages": [{"role": "user", "content": "Explain tokenization in one paragraph."}]
}'Where to get access
Developer docs
API reference, SDKs and quickstarts for MiniMax.
platform.minimax.io/document/introduction ↗