Mistral Medium 3
| Released | 2025-05-07 |
|---|---|
| Price per 1M | $0.40 in · $2.00 out |
| Overall score | 39.6 · rank 196 |
| Context window | 131,072 tokens |
| Provider | Mistral |
| Weights | Proprietary |
Scores
| Dimension | Score | Rank | Benchmarks |
|---|---|---|---|
| Overall: Balanced | 39.6 | 196 | 11 |
| Overall: Coding | 38.1 | 196 | 3 |
| Overall: Reasoning | 39.8 | 195 | 6 |
| Overall: Agentic work | 38.5 | 198 | 3 |
| Overall: Frontend | 38.2 | 181 | 3 |
| Coding | 37.5 | 184 | 1 |
| Agents | 38.2 | 163 | 1 |
| Reasoning | 41.2 | 180 | 1 |
| Math | 39.5 | 163 | 3 |
| Knowledge | 37.3 | 149 | 2 |
| General | 46.7 | 135 | 1 |
| Multimodal | 40.5 | 105 | 2 |
| Value vs price | −7.5 | 169 |
Evidence
Coding 37.5 ±5.3
| Benchmark | Result | Source | Date |
|---|---|---|---|
| LiveCodeBench (Vals run) | 44.8% | Vals AI | 2026-09-01 |
Agents 38.2 ±3.6
Reasoning 41.2 ±4.6
| Benchmark | Result | Source | Date |
|---|---|---|---|
| GPQA Diamond | 59.5% | Epoch AI | 2025-05-07 |
Math 39.5 ±4.1
| Benchmark | Result | Source | Date |
|---|---|---|---|
| FrontierMath Tiers 1-3 (2025 set) | 0.4% | Epoch AI | 2025-05-07 |
| MATH Level 5 | 81.6% | Epoch AI | 2025-05-07 |
| OTIS Mock AIME 2024-2025 | 32.2% | Epoch AI | 2025-05-07 |
Knowledge 37.3 ±1.3
| Benchmark | Result | Source | Date |
|---|---|---|---|
| Humanity's Last Exam | 4.5% | Epoch AI | 2026-11-11 |
| MMLU-Pro (Vals run) | 75.3% | Vals AI | 2026-09-01 |
General 46.7 ±4.1
| Benchmark | Result | Source | Date |
|---|---|---|---|
| LMArena Text | 1,387 | LMArena | 2026-11-09 |
Multimodal 40.5 ±2.3
| Benchmark | Result | Source | Date |
|---|---|---|---|
| MMMU-Pro (Vals run) | 63.0% | Vals AI | 2026-09-01 |
| LMArena Vision | 1,155 | LMArena | 2026-11-09 |
Results not used in the score
| Benchmark | Result | Source | Why not used |
|---|---|---|---|
| BFCL v4 | 37.6% | BFCL | A stronger configuration result was used |
Names matched to this model
mistral-medium-2505, mistral-medium-3