Mistral AIOpen weightsMistral Large 2
Best configuration ranks #491 of 610 on the BenchLeader Index at 41.2 ±3.3. Last measured 2 Sept 2026. Released 1 Nov 2024.
- Blended price
- $0.750/M
- $0.500 in · $1.50 out
- Output speed
- 47 tok/s
- First answer
- 0.68 s
- Context
- 262k
- Overall index41
- Reasoning36
- Coding40
- Agents & tools46
- Maths40
- Knowledge50
- Instruction following35
- Human preference48
- Long context26
- Composite39
Benchmark results
One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Reasoning
| Benchmark | default | Source |
|---|---|---|
| GPQA Diamond | 51.3%#197 | Epoch AI Benchmarking Hub |
| SimpleBench | 22.5%#77 | SimpleBench |
| LMArena Hard Prompts | 1320#222 | LMArena |
| GPQA Diamond (AA)not in index | 48.6%#413 | Artificial Analysis |
| Humanity's Last Exam (AA)not in index | 3.5%#507 | Artificial Analysis |
| GPQA Diamond (Vals)not in index | 68.4%#95 | Vals AI |
Coding
| Benchmark | default | Source |
|---|---|---|
| SciCode | 36.2%#124 | SciCode |
| LMArena Coding | 1354#218 | LMArena |
| LiveCodeBench | 55.3%#108 | Vals AI |
| IOI | 4.0%#48 | Vals AI |
| SWE-bench (Vals)not in index | 41.4%#81 | Vals AI |
Agents & tools
| Benchmark | default | Source |
|---|---|---|
| Terminal-Bench Hard | 6.1%#257 | Artificial Analysis |
| τ²-Bench Telecom (AA)not in index | 33.0%#238 | Artificial Analysis |
| BFCL Overall | 38.4%#36 | Berkeley Function Calling Leaderboard |
Maths
| Benchmark | default | Source |
|---|---|---|
| OTIS Mock AIME | 8.5%#211 | Epoch AI Benchmarking Hub |
| MATH Level 5 | 50.3%#52 | Epoch AI Benchmarking Hub |
| AIME (Vals) | 42.9%#64 | Vals AI |
| MGSM | 87.2%#58 | Vals AI |
Knowledge
| Benchmark | default | Source |
|---|---|---|
| AA-Omniscience | -34.4#231 | Artificial Analysis |
| MMLU-Pro | 79.8%#89 | Vals AI |
| LegalBench | 79.1%#90 | Vals AI |
| CorpFin | 61.0%#60 | Vals AI |
| TaxEval | 73.1%#53 | Vals AI |
| MedQA | 82.2%#65 | Vals AI |
Instruction following
| Benchmark | default | Source |
|---|---|---|
| IFBench | 31.6%#332 | Artificial Analysis |
Human preference
| Benchmark | default | Source |
|---|---|---|
| LMArena Text | 1314#221 | LMArena |
Long context
| Benchmark | default | Source |
|---|---|---|
| AA-LCR | 2.0%#417 | Artificial Analysis |
Composite
| Benchmark | default | Source |
|---|---|---|
| Epoch Capabilities Indexnot in index | 128.5#139 | Epoch AI Benchmarking Hub |
| AA Intelligence Index | 7.6#387 | Artificial Analysis |
Where to run it
Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.
| Provider | Speed | First token | Input $/M | Output $/M | Context | Quantisation |
|---|---|---|---|---|---|---|
| Mistral | 47 tok/s | 0.68 s | $2.00 | $6.00 | 128k | – |
What a task costs
Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0007 | – | 7.1 s |
| Summarise a 30-page report | 12,000 / 600 | $0.0069 | – | 13.4 s |
| Code edit | 6,000 / 1,500 | $0.0053 | – | 32.6 s |
| Agentic coding session | 60,000 / 4,000 | $0.036 | – | 1.4 min |
| Structured extraction | 2,000 / 200 | $0.0013 | – | 4.9 s |
See also
Data as of 9 Sept 2026. Compare with another model.