Mistral AIMistral Small 3
Mistral Small 3 is a Mistral AI open-weights model. It is not yet ranked: 3 independent results so far, and the index needs at least five across two categories. At $0.058 per million tokens blended it is among the cheapest fifth of ranked models. Output speed of 42 tokens per second puts it in the slowest quarter, with a first token in 0.3 s. Last measured 2 Sept 2026.
- Blended price
- $0.058/M
- $0.050 in · $0.080 out
- Output speed
- 42 tok/s
- OpenRouter traffic, 7-day median; not yet measured by Artificial Analysis
- First answer
- 0.32 s
- Context
- 33k
- Reasoning43
- Coding43
- Human preference43
Versions
Mistral AI has shipped 8 models under this name. Each is ranked on its own results; a newer version often has fewer results so far, which holds its index nearer the average until more arrive.
| Model | Released | Index | Rank |
|---|---|---|---|
| Mistral Small 4 | 16 Mar 2026 | 47.2 | #364 |
| Mistral Small 3.1 | 17 Mar 2025 | 39.8 | #535 |
| Mistral Small 3.1 | 17 Mar 2025 | 34.6 | #616 |
| Mistral Small 3 | 30 Jan 2025 | 38.4 | #565 |
| Mistral Small 3 | 25 Jan 2025 | – | – |
| Mistral Small 3.2 | – | 42.3 | #474 |
| Mistral Small 2402 | – | 32.0 | #623 |
| Mistral Small 3this page | – | – | – |
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 |
|---|---|---|
| LMArena Hard Prompts | 1285#252 | LMArena |
Coding
| Benchmark | default | Source |
|---|---|---|
| LMArena Coding | 1312#251 | LMArena |
Human preference
| Benchmark | default | Source |
|---|---|---|
| LMArena Text | 1274#261 | LMArena |
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 |
|---|---|---|---|---|---|---|
| DeepInfra | 42 tok/s | 0.32 s | $0.050 | $0.080 | 33k | fp8 |
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 | $0.0000 | – | 7.5 s |
| Summarise a 30-page report | 12,000 / 600 | $0.0006 | – | 14.6 s |
| Code edit | 6,000 / 1,500 | $0.0004 | – | 36.0 s |
| Agentic coding session | 60,000 / 4,000 | $0.0033 | – | 1.6 min |
| Structured extraction | 2,000 / 200 | $0.0001 | – | 5.1 s |
See also
Data as of 10 Sept 2026. Compare with another model.