MiniMax M3
MiniMax · Open Weight · rank 54 · bench-align-v5
Canonical idminimax-m3
Access typeOpen Weight
Blended $/1M$0.5275% input / 25% output
Capability shape
Seven axes from the ranking source. A missing axis is drawn as a gap.
Runtime service evidence
Measured values with the date they were observed. Nothing is inferred from a sibling model or a provider claim.
Time to first token: latency to the first answer chunk (Artificial Analysis, via BenchLM). Reasoning models include thinking time, so values can run to tens or hundreds of seconds.
Evidence key: Observed
| Measurement | Value | Observed | Last good | Evidence |
|---|---|---|---|---|
| Time to first token | 24.15 s | 2026-09-17 | 2026-09-17 | |
| Throughput | 88 tok/s | 2026-09-17 | 2026-09-17 |
Regional or per-endpoint measurements appear only when the API supplies them; none are modelled here.
Endpoint and price matrix
Every published price component, including cache reads and writes.
| Component | USD | Evidence |
|---|---|---|
| Input / 1M tokens | $0.30 | |
| Output / 1M tokens | $1.20 | |
| Cache read / 1M tokens | $0.06 | |
| Cache write / 1M tokens | Unavailable | |
| Blended / 1M (75% input / 25% output) | $0.52 |
Self-hosted listingMiniMax's official API pricing page lists the permanent 50%-off standard rate for MiniMax-M3 at $0.30 input / $0.06 cached input / $1.20 output per million tokens for prompts up to 512K. The 512K-to-1M tier is $0.60 / $0.12 / $2.40; BenchLM stores the standard short-context rate.The rates above are a hosted price matched from another provider, not a first-party list price.
Workload-aware monthly cost example
10 conversations per day × 8 messages × 22 active days, 1200 input and 400 output tokens per message, no cache. This uses the same calculator as the cost simulator, so an unavailable applicable rate makes the example unavailable too.
Benchmark record
47 matched benchmark rows with their published value, unit, and provenance.
Knowledge
| Benchmark | Value | Tasks | Difficulty | Provenance |
|---|---|---|---|---|
| Artificial Analysis Intelligence Index Artificial Analysis Intelligence Index | 29.6 | Cross-benchmark intelligence index | Display-only external reference | Artificial Analysis |
| AA-GPQA Diamond Artificial Analysis GPQA Diamond | 92.9 | Graduate-level science questions | Graduate-level science reasoning | Artificial Analysis GPQA Diamond Benchmark Leaderboard |
| AA-HLE Artificial Analysis Humanity's Last Exam | 39.0 | Expert-level questions | Frontier expert reasoning | Artificial Analysis Humanity's Last Exam Benchmark Leaderboard |
| AA-Omniscience Index Artificial Analysis Omniscience Index | 1.4 | Knowledge questions | Broad factual knowledge | AA-Omniscience: Knowledge and Hallucination Benchmark |
| AA-Omniscience Accuracy Artificial Analysis Omniscience Accuracy | 16.7 | Knowledge questions | Broad knowledge | Artificial Analysis model benchmarks |
| AA-Omniscience Hallucination Rate Artificial Analysis Omniscience Hallucination Rate | 18.4 | Knowledge questions | Factuality | Artificial Analysis model benchmarks |
| GPQA Diamond (Vals) GPQA Diamond, Vals AI run | 92.7 | Graduate-level science questions | Expert reasoning | Vals AI GPQA Diamond, Vals AI run leaderboard |
| MMLU-Pro (Vals) MMLU-Pro, Vals AI run | 84.2 | Academic multiple-choice questions | Broad academic knowledge | Vals AI MMLU-Pro, Vals AI run leaderboard |
Coding
| Benchmark | Value | Tasks | Difficulty | Provenance |
|---|---|---|---|---|
| Terminal-Bench 2.0 Terminal-Bench 2.0 | 66.0 | Terminal-based software tasks | Professional software engineering | Terminal-Bench 2.0 |
| SWE-bench Verified Software Engineering Benchmark Verified | 80.5 | 500 verified issues | Professional software engineering | SWE-bench: Can Language Models Resolve Real-World GitHub Issues? |
| SWE-bench Pro SWE-bench Pro | 59 | 1,865 repository problems | Long-horizon professional engineering | SWE-Bench Pro: Can AI Agents Solve Long-Horizon Software Engineering Tasks? |
| OpenHarmony Bench OpenHarmony Bench v1.0 | 48.4 | 153 app-development and bug-fix tasks | End-to-end OpenHarmony application development | OpenHarmony Bench: Evaluating LLMs and Coding Agents on OpenHarmony App Development |
| NL2Repo NL2Repo | 42.13 | Natural language to repository tasks | System-level software comprehension | MiniMax M2.7: Early Echoes of Self-Evolution |
| AA Coding Index Artificial Analysis Coding Index | 58.6 | Cross-benchmark coding index | Display-only external reference | Artificial Analysis model leaderboards |
| AA-SciCode Artificial Analysis SciCode | 47.1 | Scientific coding subproblems | Scientific programming | Artificial Analysis SciCode Benchmark Leaderboard |
| VIBE V2 VIBE V2 | 50.1 | End-to-end coding-agent tasks | Frontier coding-agent workflows | MiniMax M3 model card |
| SVG-Bench SVG-Bench | 63.7 | SVG generation and editing tasks | Visual coding and structured graphics generation | MiniMax M3 model card |
| KernelBench Hard KernelBench Hard | 28.8 | Hard GPU kernel coding tasks | Specialized systems programming | MiniMax M3 model card |
| LiveCodeBench (Vals) LiveCodeBench, Vals AI run | 82.2 | Competitive programming problems (easy, medium, hard) | Frontier coding | Vals AI LiveCodeBench, Vals AI run leaderboard |
| SWE-bench (Vals) SWE-bench, Vals AI run | 75.0 | Real repository issues by human time bucket | Frontier coding agents | Vals AI SWE-bench, Vals AI run leaderboard |
Mathematics
| Benchmark | Value | Tasks | Difficulty | Provenance |
|---|---|---|---|---|
| USAMO 2026 United States of America Mathematical Olympiad 2026 | 85.71 | 6 proof-based problems | International olympiad level | United States of America Mathematical Olympiad |
Reasoning
| Benchmark | Value | Tasks | Difficulty | Provenance |
|---|---|---|---|---|
| AA-LCR Artificial Analysis Long Context Reasoning | 83.0 | Long-context reasoning tasks | Long-context reasoning | Artificial Analysis model benchmarks |
| CritPt Critical Physics Tasks | 3.7 | Research-level physics questions | Research-level physics reasoning | CritPt Benchmark Leaderboard |
Instruction Following
| Benchmark | Value | Tasks | Difficulty | Provenance |
|---|---|---|---|---|
| AA-IFBench Artificial Analysis IFBench | 82.9 | Verifiable instruction constraints | Instruction precision | Artificial Analysis IFBench Benchmark Leaderboard |
Agentic
| Benchmark | Value | Tasks | Difficulty | Provenance |
|---|---|---|---|---|
| AA Harvey LAB Artificial Analysis Harvey LAB-AA | 88.4 | Legal agent tasks | Professional legal work | Artificial Analysis Harvey LAB-AA Benchmark Leaderboard |
| Terminal-Bench 2.0 Terminal-Bench 2.0 | 66 | Terminal-based software tasks | Professional software engineering | Terminal-Bench 2.0 |
| BrowseComp BrowseComp | 83.52 | Research questions requiring browsing | Hard web research | BrowseComp |
| GDPval-AA GDPval-AA | 1304 | Agentic real-world work tasks | Professional agentic workflows | DeepSeek-V4 Technical Report |
| GDPval-AA GDPval-AA normalized | 40.2 | Economically valuable tasks | Professional agentic workflows | Artificial Analysis model benchmarks |
| AA Agentic Index Artificial Analysis Agentic Index | 30.8 | Cross-benchmark agentic index | Display-only external reference | Artificial Analysis model leaderboards |
| AA-AnalystAgent Artificial Analysis AnalystAgent | 10.0 | Spreadsheet and document analysis questions | Business and data analysis | AA-AnalystAgent Benchmark Leaderboard |
| OSWorld-Verified OSWorld-Verified | 70.06 | 369 real-world computer tasks (361 when eight Google Drive tasks are excluded) | Multi-step desktop and cross-application workflows | OSWorld |
| OSWorld 2.0 OSWorld 2.0 | 4.6 | 108 long-horizon computer-use workflows | Long-horizon professional workflows | OSWorld2.0: Benchmarking Computer Use Agents on Long-Horizon Real-World Tasks |
| MCP Atlas MCP Atlas | 74.2 | Tool-integrated agent tasks | Advanced tool use | Introducing GPT-5.4 mini and nano |
| τ²-bench results τ²-Bench Tool-Agent-User Evaluation | 88.9 | Airline, retail, and telecom customer-service task sets | Dual-control customer-service workflows | τ²-Bench: Evaluating Conversational Agents in a Dual-Control Environment |
| Claw-Eval Claw-Eval | 74.5 | 300 tasks, 2,159 rubrics | Real-world general, multi-turn, and native multimodal agent execution | Claw-Eval: Towards Trustworthy Evaluation of Autonomous Agents |
| ResearchClawBench ResearchClawBench | 19.8 | 40 tasks across 10 scientific domains | Scientific research re-discovery | ResearchClawBench: A Benchmark for End-to-End Autonomous Scientific Research |
| GDPval rubrics GDPval rubrics | 74.7 | Economically valuable work tasks | Professional agentic workflows | MiniMax M3 model card |
| BankerToolBench BankerToolBench | 76.1 | Finance and banking tool-use tasks | Professional finance-agent workflows | MiniMax M3 model card |
| Terminal-Bench 2.1 (Vals) Terminal-Bench 2.1, Vals AI run | 53.6 | Difficult terminal tasks | Frontier agentic | Vals AI Terminal-Bench 2.1, Vals AI run leaderboard |
Multimodal & Grounded
| Benchmark | Value | Tasks | Difficulty | Provenance |
|---|---|---|---|---|
| MMMU-Pro Massive Multi-discipline Multimodal Understanding Pro | 78.1 | Multimodal academic reasoning | Frontier multimodal | MMMU-Pro: A More Robust Multi-discipline Multimodal Understanding Benchmark |
| AA-MMMU-Pro Artificial Analysis MMMU-Pro | 78.6 | Multimodal academic reasoning | Frontier multimodal | Artificial Analysis MMMU-Pro Benchmark Leaderboard |
| Design Arena Website Design Arena Website Elo | 1271 | Website generation comparisons | Design and website generation | OpenRouter Grok 4.3 benchmarks |
| OfficeQA Pro OfficeQA Pro | 45.1 | Document and spreadsheet tasks | Enterprise grounded reasoning | OfficeQA Pro: An Enterprise Benchmark for End-to-End Grounded Reasoning |
| OmniDocBench 1.5 OmniDocBench 1.5 | 91.6 | Document understanding tasks | Grounded document reasoning | Introducing GPT-5.4 mini and nano |
| Video-MME (with subtitle) Video-MME with subtitle | 85.4 | Video understanding | Multimodal video reasoning | Qwen3.6 launch benchmarks |
| VideoMMMU VideoMMMU | 84.6 | Video-grounded expert reasoning | Frontier multimodal video reasoning | Qwen3.6 launch benchmarks |
Lifecycle and limitations log
Lifecycle events the source associates with this model, plus what this profile does not claim.
No lifecycle event references this model.
That is not evidence the model has no lifecycle plan — only that this source published none.
What this profile does not claimValues are reproduced exactly as their sources published them, in the units those sources declared; none are converted, interpolated, or averaged across providers. Any field marked unavailable was attempted and not returned — the attempt timestamp is in each badge. Last attempted fetch for this model’s score: 2026-09-17 17:17 UTC.