Mistral Medium 3.5
| Released | 2026-04-30 |
|---|---|
| Price per 1M | $1.50 in · $7.50 out |
| Overall score | 46.9 · rank 139 |
| Context window | 262,144 tokens |
| Provider | Mistral |
| Weights | Proprietary |
Scores
| Dimension | Score | Rank | Benchmarks |
|---|---|---|---|
| Overall: Balanced | 46.9 | 139 | 16 |
| Overall: Coding | 47.4 | 128 | 8 |
| Overall: Reasoning | 43.7 | 168 | 6 |
| Overall: Agentic work | 48.4 | 127 | 6 |
| Overall: Frontend | 41.4 | 169 | 4 |
| Coding | 51.7 | 107 | 1 |
| Agents | 49.5 | 108 | 2 |
| Reasoning | 38.4 | 203 | 3 |
| Math | 53.3 | 85 | 1 |
| Knowledge | 38.9 | 145 | 2 |
| Frontend | 35.1 | 123 | 2 |
| General | 49.1 | 116 | 4 |
| Multimodal | 48.2 | 80 | 1 |
| Value vs price | −4.2 | 154 |
Configurations
| Configuration | Balanced | Benchmarks |
|---|---|---|
| Unspecified setting | 43.2 ±5.4 | 6 |
| High | 47.0 ±1.5 | 10 |
Evidence
Coding 51.7 ±4.2
| Benchmark | Result | Source | Date |
|---|---|---|---|
| SWE-bench Verified (Vals run) | 66.4% | Vals AI | 2026-09-01 |
Agents 49.5 ±2.6
| Benchmark | Result | Source | Date |
|---|---|---|---|
| Terminal-Bench 2.1 (Vals run) | 39.0% | Vals AI | 2026-10-01 |
| Tax Agent Bench | 4.5% | Vals AI | 2026-11-09 |
Reasoning 38.4 ±6.8
| Benchmark | Result | Source | Date |
|---|---|---|---|
| Extended NYT Connections | 12.9% | Lech Mazur | 2026-11-10 |
| GPQA Diamond (Vals run) | 34.8% | Vals AI | 2026-09-01 |
| Kagi Reasoning Benchmark | 41.4% | Kagi | 2026-11-11 |
Math 53.3 ±2.7
| Benchmark | Result | Source | Date |
|---|---|---|---|
| ProofBench v1.1 | 9.0% | Vals AI | 2026-11-09 |
Knowledge 38.9 ±3.9
| Benchmark | Result | Source | Date |
|---|---|---|---|
| MMLU-Pro (Vals run) | 75.3% | Vals AI | 2026-09-01 |
| Legal Research Bench | 9.1% | Vals AI | 2026-11-09 |
Frontend 35.1 ±0.7
| Benchmark | Result | Source | Date |
|---|---|---|---|
| Vibe Code Bench v1.1 | 2.9% | Vals AI | 2026-11-09 |
| LMArena WebDev | 1,263 | LMArena | 2026-11-09 |
General 49.1 ±2.4
| Benchmark | Result | Source | Date |
|---|---|---|---|
| LLM Debate Benchmark | 1,333 | Lech Mazur | 2026-10-26 |
| EQ-Bench 4 | 983 | EQ-Bench | 2026-10-24 |
| LMArena Text | 1,427 | LMArena | 2026-11-09 |
| Creative Story-Writing (Lech Mazur) | 16.0% | Lech Mazur | 2026-10-23 |
Multimodal 48.2 ±4.1
| Benchmark | Result | Source | Date |
|---|---|---|---|
| LMArena Vision | 1,199 | LMArena | 2026-11-09 |