Mistral AIMistral Small 4
Best configuration ranks #357 of 610 on the BenchLeader Index at 47.1 ±3.3. Released 16 Mar 2026.
- Blended price
- $0.262/M
- $0.150 in · $0.600 out
- Output speed
- 168 tok/s
- First answer
- 13 s
- Context
- 256k
- Overall index47
- Reasoning39
- Agents & tools49
- Knowledge50
- Instruction following49
- Multimodal41
- Long context51
- Composite43
Reasoning-effort configurations
The same model behaves differently depending on how much it is allowed to think. Each row is one setting, scored only on the benchmarks that were run at that setting. “Default” means the publisher did not say which setting was used.
| Effort | Index | Rank | Speed | First answer | Chat reply cost | Categories |
|---|---|---|---|---|---|---|
| no reasoning | 42.2 | #466 | 153 tok/s | 0.73 s | $0.0002 | Agents & tools 43 · Composite 40 · Instruction following 36 · Knowledge 41 · Long context 39 · Multimodal 30 |
| defaultbest | 47.1 | #357 | 168 tok/s | 13 s | $0.0002 | Agents & tools 49 · Composite 43 · Instruction following 49 · Knowledge 50 · Long context 51 · Multimodal 41 · Reasoning 39 |
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 | no reasoning | default | Source |
|---|---|---|---|
| GPQA Diamond (AA)not in index | 57.1%#376 | 76.9%#219 | Artificial Analysis |
| Humanity's Last Exam (AA)not in index | 3.8%#479 | 9.9%#260 | Artificial Analysis |
| Kagi LLM Benchmark | – | 40.5%#102 | Kagi LLM Benchmark |
Coding
| Benchmark | no reasoning | default | Source |
|---|---|---|---|
| SciCode (AA)not in index | – | 38.8%#122 | Artificial Analysis |
Agents & tools
| Benchmark | no reasoning | default | Source |
|---|---|---|---|
| Terminal-Bench Hard | 10.6%#213 | 17.4%#173 | Artificial Analysis |
| τ²-Bench Telecom (AA)not in index | 18.4%#332 | 41.2%#214 | Artificial Analysis |
Knowledge
| Benchmark | no reasoning | default | Source |
|---|---|---|---|
| AA-Omniscience | -48.5#303 | -30.4#217 | Artificial Analysis |
Instruction following
| Benchmark | no reasoning | default | Source |
|---|---|---|---|
| IFBench | 32.8%#321 | 48.2%#182 | Artificial Analysis |
Multimodal
| Benchmark | no reasoning | default | Source |
|---|---|---|---|
| MMMU-Pro | 46.5%#216 | 56.8%#188 | Artificial Analysis |
Long context
| Benchmark | no reasoning | default | Source |
|---|---|---|---|
| AA-LCR | 28.3%#326 | 49.7%#256 | Artificial Analysis |
Composite
| Benchmark | no reasoning | default | Source |
|---|---|---|---|
| AA Intelligence Index | 9.0#342 | 11.4#286 | 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 | – | 14.4 s |
| Summarise a 30-page report | 12,000 / 600 | $0.0022 | – | 16.2 s |
| Code edit | 6,000 / 1,500 | $0.0018 | – | 21.5 s |
| Agentic coding session | 60,000 / 4,000 | $0.011 | – | 36.4 s |
| Structured extraction | 2,000 / 200 | $0.0004 | – | 13.8 s |
See also
Data as of 9 Sept 2026. Compare these configurations.