Mistral AIMistral 7B v0.3
Best configuration ranks #608 of 610 on the BenchLeader Index at 31.5 ±3.9. Last measured 2 Sept 2026. Released 27 May 2024.
- Blended price
- $0.250/M
- $0.250 in · $0.250 out
- Output speed
- 78 tok/s
- First answer
- 0.75 s
- first token 0.75 s
- Context
- 66k
- Overall index32
- Reasoning20
- Coding29
- Maths25
- Knowledge26
- Instruction following25
- Human preference28
- Long context25
- Composite35
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 | Trend |
|---|---|---|---|
| GPQA Diamond | 15.2%#268 | Epoch AI Benchmarking Hub | |
| LMArena Hard Prompts | 1156#312 | LMArena | |
| GPQA Diamond (AA)not in index | 17.7%#523 | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | 4.5%#413 | Artificial Analysis |
Coding
| Benchmark | default | Source | Trend |
|---|---|---|---|
| LMArena Coding | 1186#309 | LMArena |
Agents & tools
| Benchmark | default | Source | Trend |
|---|---|---|---|
| τ²-Bench Telecom (AA)not in index | 0.0%#375 | Artificial Analysis |
Maths
| Benchmark | default | Source | Trend |
|---|---|---|---|
| OTIS Mock AIME | 0.3%#253 | Epoch AI Benchmarking Hub | |
| MATH Level 5 | 3.6%#89 | Epoch AI Benchmarking Hub |
Knowledge
| Benchmark | default | Source | Trend |
|---|---|---|---|
| AA-Omniscience | -64.9#421 | Artificial Analysis | |
| LegalBench | 53.8%#131 | Vals AI |
Instruction following
| Benchmark | default | Source | Trend |
|---|---|---|---|
| IFBench | 19.9%#393 | Artificial Analysis |
Human preference
| Benchmark | default | Source | Trend |
|---|---|---|---|
| LMArena Text | 1149#314 | LMArena |
Long context
| Benchmark | default | Source | Trend |
|---|---|---|---|
| AA-LCR | 0.0%#419 | Artificial Analysis |
Composite
| Benchmark | default | Source | Trend |
|---|---|---|---|
| Epoch Capabilities Indexnot in index | 112.0#185 | Epoch AI Benchmarking Hub | |
| AA Intelligence Index | 5.0#517 | Artificial Analysis |
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 | $0.0002 | – | 4.6 s |
| Summarise a 30-page report | 12,000 / 600 | $0.0032 | – | 8.4 s |
| Code edit | 6,000 / 1,500 | $0.0019 | – | 19.9 s |
| Agentic coding session | 60,000 / 4,000 | $0.016 | – | 51.9 s |
| Structured extraction | 2,000 / 200 | $0.0006 | – | 3.3 s |
See also
Data as of 9 Sept 2026. Compare with another model.