MiniMax M3
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,...
MiniMax M3 costs $0.23 per 1M input tokens and $0.96 per 1M output tokens at the standard tier, the lowest default-region price among the 13 providers SF·SIX tracks, from CoreWeave (fp4). On MiniMax’s own API it costs $0.30 / $1.20. Cached input reads cost $0.05 per 1M tokens. Updated ; prices checked .
- Developer
- MiniMax
- Input modalities
- text, image, video
- Output modalities
- text
- Context tokens
- 1,048,576
- Latest price observation
- 2026-10-17
- Lowest paired standard token quote
- $0.23 input / $0.96 output per 1M tokens
Dated price and provider records
| Provider and channel | Input | Output or native rate | Unit and status | Source and observation |
|---|---|---|---|---|
| CoreWeaveOpenRouter | $0.23 | $0.96 | 1M tokenspriced · USD | OpenRouter2026-10-17 |
| GMI CloudOpenRouter | $0.24 | $0.96 | 1M tokenspriced · USD | OpenRouter2026-10-17 |
| DeepInfraOpenRouter | $0.28 | $1.10 | 1M tokenspriced · USD | OpenRouter2026-10-17 |
| AtlasCloudOpenRouter | $0.30 | $1.20 | 1M tokenspriced · USD | OpenRouter2026-10-17 |
| MiniMaxOpenRouter | $0.30 | $1.20 | 1M tokenspriced · USD | OpenRouter2026-10-17 |
| Novita AIOpenRouter | $0.30 | $1.20 | 1M tokenspriced · USD | OpenRouter2026-10-17 |
| ParasailOpenRouter | $0.30 | $1.20 | 1M tokenspriced · USD | OpenRouter2026-10-17 |
| StreamLakeOpenRouter | $0.30 | $1.20 | 1M tokenspriced · USD | OpenRouter2026-10-17 |
| Together AIOpenRouter | $0.30 | $1.20 | 1M tokenspriced · USD | OpenRouter2026-10-17 |
| Venice AIOpenRouter | $0.30 | $1.20 | 1M tokenspriced · USD | OpenRouter2026-10-17 |
| ModelRunOpenRouter | $0.30 | $1.30 | 1M tokenspriced · USD | OpenRouter2026-10-17 |
| MaraOpenRouter | $0.60 | $2.40 | 1M tokenspriced · USD | OpenRouter2026-10-17 |
| SambaNovaOpenRouter | $0.60 | $2.40 | 1M tokenspriced · USD | OpenRouter2026-10-17 |
| DeepInfraOpenRouter · Flex tier | $0.224 | $0.88 | 1M tokenspriced · USD | OpenRouter2026-10-17 |
Prices by service tier
| Tier | Input / 1M | Output / 1M | Cache read / 1M | Cheapest provider | vs standard (blended) |
|---|---|---|---|---|---|
| Standard | $0.23 | $0.96 | $0.05 | CoreWeave | 100% |
| Flex | $0.224 | $0.88 | $0.0448 | DeepInfra | 94% |
Each tier’s cheapest endpoint. Standard is the default-region headline; flex and batch trade speed or latency for a lower price, priority and fast cost more for faster service.
Speed by provider
| Provider and terms | Throughput | Latency | Uptime | Input / 1M |
|---|---|---|---|---|
| CoreWeavefp4 | 144 tok/s | 469 ms | 100.0% | $0.23 |
| Mara | 118 tok/s | 1.34 s | Not reported | $0.60 |
| Parasailfp8 | 107 tok/s | 1.02 s | 100.0% | $0.30 |
| ModelRunfp8 | 100 tok/s | 506 ms | 100.0% | $0.30 |
| AtlasCloudfp8 | 96 tok/s | 1.34 s | 100.0% | $0.30 |
| Together AI | 93 tok/s | 501 ms | 100.0% | $0.30 |
| MiniMaxfp8 | 86 tok/s | 1.15 s | 100.0% | $0.30 |
| Novita AIfp8 | 83 tok/s | 2.53 s | 100.0% | $0.30 |
| StreamLakefp8 | 68 tok/s | 1.34 s | Not reported | $0.30 |
| GMI Cloudfp8 | 52 tok/s | 2.71 s | 100.0% | $0.24 |
| Venice AIfp8 | 33 tok/s | 875 ms | Not reported | $0.30 |
| DeepInfraFlex tierfp8 | 33 tok/s | 2.35 s | Not reported | $0.224 |
Previous 30 minutes; reported by OpenRouter. Fastest first.
Usage on OpenRouter
| UTC day | Tokens |
|---|---|
| 2026-10-10 | 179,003,985,096 |
| 2026-10-11 | 169,862,382,976 |
| 2026-10-12 | 202,523,712,936 |
| 2026-10-13 | 214,019,237,814 |
| 2026-10-14 | 205,834,854,036 |
| 2026-10-15 | 208,245,074,899 |
| 2026-10-16 | 210,130,855,533 |
OpenRouter rankings (CC BY 4.0), named daily rows. Most-used AI models
Benchmarks
Artificial Analysis Intelligence Index 29.2, coding index 58.6, agentic index 29.5 (Artificial Analysis via OpenRouter, CC BY 4.0; not an SF·SIX measurement).
Prediction markets
MiniMax M3 meets the rule for SF·SIX price and usage ladders; the current average standard input price across qualifying OpenRouter endpoints is $0.3464 per 1M tokens. Open MiniMax M3 markets (Hedera testnet, test funds only) · How AI prediction markets work
Compare and estimate
Sources and method
Prices are USD per 1M tokens at the standard service tier in the default region: the cheapest qualifying provider per model, with flex, batch, priority and fast endpoints and unusually low outliers never used as the headline. “Blended” weights input and output 3:1. Prices from OpenRouter; each row below links its source and observation date. Methodology and editorial policy.
Frequently asked questions
How much does MiniMax M3 cost per 1M tokens?
$0.23 per 1M input tokens and $0.96 per 1M output tokens at the standard tier, the lowest default-region price SF·SIX tracks, from CoreWeave (fp4). On MiniMax’s own API it costs $0.30 / $1.20.
Which providers offer MiniMax M3?
13 providers with 14 priced endpoints: AtlasCloud, CoreWeave, DeepInfra, GMI Cloud, Mara, MiniMax, ModelRun, Novita AI and others.
What is the context window of MiniMax M3?
1,048,576 tokens.
How fast is MiniMax M3?
The fastest measured endpoint is CoreWeave (fp4) at 144 tok/s, as reported by OpenRouter for the previous 30 minutes.
How much is MiniMax M3 used?
210,130,855,533 tokens on 2026-10-16 (UTC) through OpenRouter, per OpenRouter’s public rankings. This is OpenRouter traffic only, not all usage.
These are dated, non-executable observations. Verify the exact endpoint, price and terms with the source before use. Unknown values are not free.