MiniMax multimodal model for long-context coding, perception, and agent planning
MiniMax-M3 is offered by 53 providers on this page. Public prices are shown for 47 of them. Its official list price is $0.30 per 1M input tokens and $1.20 per 1M output tokens. The lowest paid channel is Vultr at $0.20 / $0.90 per 1M, about 3.6× below the list price. It also has 2 free ($0) channels and 4 covered by a paid subscription; free tiers usually carry rate limits, and subscription-covered access bills $0 per token only after the subscription fee.
Artificial Analysis rates it 29.2 on the Intelligence Index, with 58.6 for coding and 29.5 for agentic tasks. Against its lowest blended price of $0.375 per 1M, that is roughly 78 index points per dollar, which is the value ratio the leaderboards rank on. Median output speed is 97 tokens per second, with 2.03s to the first token. Running one task of the Artificial Analysis suite costs about $0.5076, which reflects how many tokens its reasoning consumes rather than the unit price alone.
The context window is 1,000,000 tokens at the reference host, but hosts report different limits, from 262,144 to 1,048,576, so the usable window depends on the provider you pick. It supports reasoning, tool use, and structured output. Accepted input modalities are Text, Image, Video, and PDF. The weights are open, so it can also be self-hosted or served through a gateway of your choice. Providers report the capability flags inconsistently, so verify a specific feature against the host you plan to use. Its training knowledge cuts off at 2025-01.
Quality is independently evaluated by Artificial Analysis. Speed/latency are model-level medians.
| Provider | Tier | Input | Output | Cache read | Cache write | Context | Output limit | Status |
|---|---|---|---|---|---|---|---|---|
| MiniMax Token Plan (minimax.io) MiniMax-M3 | Gateway | Subscription | $0 | $0 | 1,048,576 ⚠ | 512,000 | ||
| MiniMax Token Plan (minimax.cn) MiniMax-M3 | Gateway | Subscription | $0 | $0 | 1,048,576 ⚠ | 512,000 | ||
| Volcengine Ark Coding Plan | Gateway | Subscription | $0 | — | 1,048,576 ⚠ | 512,000 | ||
| SCNet Token Plan MiniMax-M3 | Gateway | Subscription | $0 | — | 1,048,576 ⚠ | 512,000 | ||
| Nvidia | Cloud | Free | — | — | 1,000,000 | 16,384 | deprecated | |
| Kenari | Gateway | Free | — | — | 1,048,576 ⚠ | 512,000 | ||
| Vultr | Cloud | $0.20 | $0.90 | — | — | 524,288 ⚠ | 512,000 | |
| EmpirioLabs AI | Gateway | $0.23 tiered >512K: $0.45 | $0.90 | $0.045 | — | 1,000,000 | 524,288 | |
| CoreWeave MiniMaxAI/MiniMax-M3 | Cloud | $0.23 | $0.96 | $0.05 | — | 262,144 ⚠ | 262,144 | |
| Vancine MiniMax-M3 | Gateway | $0.24 | $0.96 | $0.048 | — | 1,048,576 ⚠ | 512,000 | |
| MiniMax (minimax.cn)Official MiniMax-M3 | First-party | $0.30 tiered >512K: $0.60 | $1.20 | $0.06 | — | 1,000,000 | 512,000 | |
Sorted by blended price (input×0.75 + output×0.25) asc. The official channel always shows regardless of rank. Whether a gateway's low price is actually usable can't be verified.
Interleaved thinking (reasoning between tool calls) is declared by 12 of 53 providers.
15 / 53 providers expose no reasoning control (reasoning_options: []).
| Mode | Provider | Input | Output | Cache read | Cache write |
|---|---|---|---|---|---|
| priority | Fireworks AI | $0.45 | $1.80 | $0.09 | — |
6 more channels offer $0 (MiniMax Token Plan (minimax.io), MiniMax Token Plan (minimax.cn), Volcengine Ark Coding Plan etc.); free tiers usually have rate limits and no SLA, excluded from ranking.
| Name | Conditions | Score | Metric | Source |
|---|---|---|---|---|
| SWE-Bench Verified | harness: Claude Code | 80.5 | resolved | Source ↗ |
| SWE-Bench Pro | harness: Claude Code | 59 | resolve rate | Source ↗ |
| Terminal-Bench | v2.1 | 66 | success rate | Source ↗ |
| BrowseComp | — | 83.52 | accuracy | Source ↗ |
| MCP Atlas | — | 74.2 | success rate | Source ↗ |
| OSWorld-Verified | — | 70.06 | success rate | Source ↗ |
The same benchmark scores very differently across harness / dataset, so the qualifying conditions must be shown together.
Individual evaluations run by Artificial Analysis, on the same reasoning tier as the intelligence score above. Each benchmark has its own task set and harness, so rows are not comparable with one another. The Intelligence Index above draws on a different, newer set of evaluations.