MiniMax-M3
MiniMax-M3 is a reasoning model from MiniMax released on May 31, 2026. It reads text, images and video and writes text, with a 1M token context window and up to 512K tokens per reply.
Show moreShow less
It costs $0.30 per million input tokens and $1.20 per million output tokens. Cached input bills at $0.06, 80% less than fresh input.
It supports tool calling, structured JSON output and prompt caching, and its weights are open, so you can host it yourself.
43 providers serve it, including Deep Infra, Fireworks AI and Hugging Face. The lowest input price is $0.20 per million tokens at Vultr.
On the Artificial Analysis Intelligence Index it scores 29.2, which places it at number 24 among the 43 models I track with independent scores.
- Modalities
- In / out price
- $0.30 / $1.20 per 1M
- Context
- 1M
- Max output
- 512K
- Released
- May 31, 2026
- Knowledge cutoff
- n/a
- Tool calling
- Structured output
- Reasoning
- Prompt caching
- Open weights
- Free to use
MiniMax-M3 pricing
USD, as listed on Oct 15, 2026. Most bills are input heavy, so the cached input rate matters as much as the headline price.
| Input | $0.30 / 1M tokens | |
|---|---|---|
| Output | $1.20 / 1M tokens | |
| Cached input (read) | $0.06 / 1M tokens | 80% off input |
| Prompts over 200K | $0.60 in, $2.40 out | long context rate |
| Prompts over 512K | $0.60 in, $2.40 out | long context rate |
MiniMax-M3 providers
43 providers serve it, including Deep Infra, Fireworks AI and Hugging Face. The lowest input price is $0.20 per million tokens at Vultr. Each host sets its own price and limits, so the cheapest one is not always the one with the longest context.
| Provider | Input /M | Output /M | Cache read /M | Context | Max output |
|---|---|---|---|---|---|
| MiniMax (minimax.io) First party | $0.30 | $1.20 | $0.06 | 1M | 512K |
| Deep Infra | $0.28 | $1.10 | $0.056 | 524K | 512K |
| Fireworks AI | $0.30 | $1.20 | $0.06 | 512K | 512K |
| Hugging Face | $0.30 | $1.20 | – | 524K | 512K |
| Nebius Token Factory | $0.30 | $1.20 | – | 1.05M | 1.05M |
| Novita AI | $0.30 | $1.20 | $0.06 | 1M | 131K |
| SiliconFlow | $0.30 | $1.20 | $0.06 | 1.05M | 131K |
| Vercel AI Gateway | $0.30 | $1.20 | $0.06 | 512K | 512K |
| SambaNova | $0.60 | $2.40 | $0.06 | 1.05M | 1.05M |
| Vultr | $0.20 | $0.90 | – | 524K | 512K |
| EmpirioLabs AI | $0.225 | $0.90 | $0.045 | 1M | 524K |
| CoreWeave | $0.23 | $0.96 | $0.05 | 262K | 262K |
Showing 12 of 43 providers: the lab first, then the major clouds, then the cheapest resellers.
What MiniMax-M3 costs per month
Pick a workload close to yours or type your own volumes.
Independent benchmarks
Measured by third parties, not reported by MiniMax. Each links to its source and shows the date I checked it.
A composite of 10 independent evaluations (agents 30%, coding 20%, scientific reasoning 20%, general 30%), run by Artificial Analysis on each model, version 4.3.2.
Artificial Analysis, checked Oct 10, 2026
Human preference: people compare two anonymous answers and pick the better one. Ratings from the Text Arena overall board, style control on.
Arena (formerly LMArena), checked Oct 10, 2026
Median output tokens per second measured by Artificial Analysis on the provider API, for the same model setting as its index score.
Artificial Analysis, checked Oct 10, 2026
Benchmarks MiniMax reports (8)
- swe-bench-verified (SWE-Bench Verified, internal infrastructure, Claude Code scaffolding, average of 4 runs)80.5%
- swe-bench-pro (SWE-Bench Pro, internal infrastructure, Claude Code scaffolding)59%
- terminal-bench-2.1 (Terminal Bench 2.1, Terminus 2 scaffolding, 2h timeout, max output 128K)66%
- browsecomp (BrowseComp, WebExplorer agent framework, history discarded past 64K tokens)83.5%
- osworld-verified (OSWorld-Verified, 361 samples, relative coordinates)75.2%
- mmmu-pro (MMMU-Pro)78.1%
- mcp-atlas (MCP Atlas, public set, Gemini 2.5 Pro scorer)74.2%
- posttrainbench (PostTrainBench, Claude Code, 12 hours)37.1
Labs pick their own settings and harnesses, so compare these only within one lab.
Supported request parameters
What you can set when you call MiniMax-M3 through an OpenAI compatible API.
- Tools
tools - Structured outputs
structured_outputs - Reasoning
reasoning - Temperature
temperature
Alternatives to MiniMax-M3
Models that take the same inputs and call tools, closest in price, cheaper ones first.
- 37K context
Reads text, images, audio and video. Tool calling, structured output, reasoning. Served by 2 providers.
by perceptronSep 25, 2026$0.15/M input tokens$1.50/M output tokens2 providers
- 128K context
Reads text, images and video. Tool calling, structured output, reasoning. Served by 13 providers.
by z-aiDec 8, 2025$0.30/M input tokens$0.90/M output tokens13 providers
- 66K context
Reads text, images, audio and video. Tool calling, structured output, reasoning.
by qwenSep 22, 2025$0.25/M input tokens$0.97/M output tokens
- 1M context
Reads text, images and video. Tool calling, structured output, reasoning. Served by 17 providers.
by qwenApr 2, 2026$0.1875/M input tokens$1.13/M output tokens17 providers
- 256K context
Reads text, images and video. Tool calling, structured output, reasoning. Served by 19 providers.
by stepfunMay 29, 2026$0.185/M input tokens$1.11/M output tokens19 providers
- Index 46.3
Reads text, images, audio and video. Tool calling, structured output, reasoning. Served by 17 providers.
by xiaomiSep 22, 20261.05M context$0.435/M input tokens$0.87/M output tokens17 providers
More from MiniMax
MiniMax H3 Max
VideoGenerates video from text and images.
by minimaxAug 27, 2026
MiniMax H3
VideoGenerates video from text and images.
by minimaxJul 31, 2026
- 512K context
Reads text and images. Tool calling, structured output, reasoning. Served by 2 providers.
by minimaxMay 3, 2026$0.30/M input tokens$1.20/M output tokens2 providers
- 205K context
Reads text. Tool calling, structured output, reasoning. Served by 35 providers.
by minimaxMar 18, 2026$0.30/M input tokens$1.20/M output tokens35 providers
- 205K context
Reads text. Tool calling, structured output, reasoning. Served by 16 providers.
by minimaxMar 18, 2026$0.60/M input tokens$2.40/M output tokens16 providers
- 205K context
Reads text. Tool calling, structured output, reasoning.
by minimaxMar 18, 2026$0.60/M input tokens$2.40/M output tokens
MiniMax-M3 questions
How much does the MiniMax-M3 API cost?
$0.30 per million input tokens and $1.20 per million output tokens, with cached input at $0.06. Prompts over 200K tokens bill at $0.60 input and $2.40 output. The estimator on this page turns that into a monthly bill for your own workload.
What is the context window of MiniMax-M3?
1,000,000 tokens of input, and it can write up to 512,000 tokens in one reply. As a rough guide, a million tokens is around 750,000 English words.
Which providers serve MiniMax-M3?
43 providers serve it, including Deep Infra, Fireworks AI and Hugging Face. The lowest input price is $0.20 per million tokens at Vultr. Prices differ by host, so the Providers table on this page lists each one with its input and output price.
Does MiniMax-M3 support tool calling?
Yes. It accepts tool definitions and can return JSON that matches a schema, which is what an agent needs to call your APIs.
Is MiniMax-M3 open source?
Its weights are published, so you can run it on your own hardware or a cloud of your choice. Read the license before commercial use.
When was MiniMax-M3 released?
May 31, 2026.
How good is MiniMax-M3 compared with other models?
It scores 29.2 on the Artificial Analysis Intelligence Index and 1440 on the Arena text leaderboard. Benchmarks are a shortlist, not a verdict: test the top two or three candidates on your own tasks before you commit.
Catalog data from models.dev (MIT licensed), refreshed daily, last on Oct 15, 2026. Independent scores from my hand checked rankings, sources linked above. The summary is written from these facts.
Building on MiniMax-M3?
I test your real tasks on it and the two closest alternatives, then ship the agent on whichever wins on cost and accuracy.
Book Meeting