token.app › MiniMax › MiniMax M3
MiniMax M3 pricing & benchmarks
MiniMax M3 is available from MiniMax at $0.300 per million input tokens and $1.20 per million output tokens ($0.525 blended at 3:1). It accepts up to 1,048,576 tokens of context. MiniMax-M3 is a multimodal foundation model from MiniMax. It supports text, image, and video inputs with text output, a 1M-token context window, and is suited for long-horizon agentic work, coding,...
Pricing
Per million tokens · OpenRouter
Input$0.300
Output$1.20
Blended (3:1)$0.525
Specification
minimax/minimax-m3
Context window1.0M
Max output512K
ReleasedMay 2026
LicenceOpen weights
Modalitiesin:text, in:image, in:video, out:text
Benchmarks
Independently run — not vendor self-reports. Every score links its source.
| Benchmark | Score | Config | Run | Source |
|---|---|---|---|---|
| GPQA Diamond | 90.9% ±2.0 | thinking effort | 2026-08-10 | Independently run · Epoch AI |
| OTIS Mock AIME | 71.1% ±6.8 | thinking effort | 2026-08-10 | Independently run · Epoch AI |
Scores published by Epoch AI, 'AI Benchmarking Hub', used under CC BY 4.0.
Cheaper models that score at least as well on GPQA Diamond
Strictly lower blended price AND an equal-or-higher GPQA Diamond score. One benchmark is one dimension — a model that wins here may still be weaker at your specific task.
| Model | Provider | Blended $/1M | GPQA Diamond | |
|---|---|---|---|---|
| GPT-5.6 Luna | OpenAI | $0.450 (−14%) | 91.6% | Compare → |
Where to run MiniMax M3
14 options across 13 providers · 2.7× between cheapest and dearest. Blended at 3:1. ⚠ marks an option serving less context than the best available, or re-pricing above a prompt-length threshold. Bracketed labels are the provider's own service tier or region.
| Host | Blended $/1M | In / Out | Context | Weights | Uptime 24h |
|---|---|---|---|---|---|
| DeepInfra (fp8/flex) ⚠ 524K context, not 1.0M |
$0.388 | $0.224 / $0.880 | 524K | fp8 | 100.0% |
| CoreWeave ⚠ 262K context, not 1.0M |
$0.412 | $0.230 / $0.960 | 262K | fp4 | 99.7% |
| GMICloud | $0.420 | $0.240 / $0.960 | 1.0M | fp8 | 99.5% |
| DeepInfra ⚠ 524K context, not 1.0M |
$0.485 | $0.280 / $1.10 | 524K | fp8 | 99.8% |
| StreamLake | $0.525 | $0.300 / $1.20 | 1M | fp8 | 98.5% |
| Venice ⚠ 524K context, not 1.0M |
$0.525 | $0.300 / $1.20 | 524K | fp8 | 96.6% |
| Together ⚠ 524K context, not 1.0M |
$0.525 | $0.300 / $1.20 | 524K | — | 99.4% |
| Parasail | $0.525 | $0.300 / $1.20 | 1.0M | fp8 | 99.9% |
| AtlasCloud ⚠ 524K context, not 1.0M |
$0.525 | $0.300 / $1.20 | 524K | fp8 | 99.1% |
| Novita | $0.525 | $0.300 / $1.20 | 1M | fp8 | 100.0% |
| Minimax ⚠ 524K context, not 1.0M |
$0.525 | $0.300 / $1.20 | 524K | fp8 | 99.9% |
| ModelRun | $0.550 | $0.300 / $1.30 | 1.0M | fp8 | 99.8% |
| Mara | $1.05 | $0.600 / $2.40 | 1.0M | — | 85.7% |
| SambaNova | $1.05 | $0.600 / $2.40 | 1.0M | — | 94.3% |
Host prices and uptime from OpenRouter, synced 2026-10-22. Lower-precision weights (fp4, fp8) can cost less and score worse than the same model at bf16 — the score above is not host-specific.
Compare MiniMax M3
FAQ
How much does MiniMax M3 cost?
MiniMax M3 costs $0.300 per million input tokens and $1.20 per million output tokens on OpenRouter. At a 3:1 input:output mix that blends to $0.525 per million tokens.
What is MiniMax M3's context window?
MiniMax M3 accepts up to 1,048,576 tokens of context, and can generate up to 512,000 output tokens.
How good is MiniMax M3 on benchmarks?
MiniMax M3 scores 90.9% on GPQA Diamond, independently run and published by Epoch AI. 2 benchmark results are listed on this page, each with its run date and source.
Is there a cheaper model as good as MiniMax M3?
Yes — 1 model in our catalogue cost less per token than MiniMax M3 and score at least as high on the same benchmark. The cheapest is GPT-5.6 Luna at $0.450 per million tokens blended.