Mistral Large 4 Preview
Mistral Large 4 Preview places in the top three on SAFE-Teen Relationship boundaries.
34
Benchmarks with a score
0
Ranked first, out of five or more models
1
In the top three
Mistral AI Oct 2026 Mistral Large 4 Preview
Open weights
152.7
#57 of 256
Epoch Capabilities Index
- GPQA Diamond 88.1% #64 /229
- Harvey LAB (Legal Agent Benchmark) 15.8% #10 /80
- Vals Finance Agent v2 54.7% #26 /80
Where it ranks
Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.
Knowledge and language
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| SimpleQA Verified | 20% | #81 of 94 | Epoch AI run |
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| OTIS Mock AIME 2024-2025 | 98.3% | #33 of 212 | Epoch AI run | |
| GPQA Diamond | 88.1% | #64 of 229 | Epoch AI run | |
| FrontierMath Tiers 1-3 v2 | 58.6% | #47 of 92 | Epoch AI run | |
| FrontierMath Tier 4 v2 | 26.8% | #46 of 74 | Epoch AI run | |
| BioMysteryBench | 67% | #22 of 30 | Official leaderboard | |
| ProofBench v1.1 | 10% | #47 of 52 | Official leaderboard | |
| MysteryMechanism | 12.2% | #31 of 31 | Official leaderboard |
Coding
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Vibe Code Bench | 78.4% | #37 of 120 | Official leaderboard | |
| Code Migration | 30.6% | #42 of 78 | Official leaderboard | |
| IOI v2 (Vals) | 45.3% | #32 of 45 | Official leaderboard | |
| MirrorCode | 27.8% | #31 of 39 | Epoch AI run | |
| Vibe Code Bench 1-100 | 14.3% | #21 of 26 | Official leaderboard |
Agents and tools
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Chess Puzzles | 33% | #51 of 93 | Epoch AI run | |
| Mystery Game Puzzles | 17% | #63 of 84 | Epoch AI run |
Professional work
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Harvey LAB (Legal Agent Benchmark) | 15.8% | #10 of 80 | Official leaderboard | |
| Harvey LAB v1.1 (HLAB-Vals-1.1) | 17.5% | #8 of 38 | Official leaderboard | |
| Vals Finance Agent v2 | 54.7% | #26 of 80 | Official leaderboard | |
| Tax Agent Bench | 30.4% | #34 of 84 | Official leaderboard | |
| APEX-Agents | 51.6% | #32 of 61 | Collected by Epoch AI, source not named | |
| MedScribe | 80.4% | #53 of 97 | Official leaderboard | |
| Legal Research Bench | 31.7% | #53 of 93 | Official leaderboard | |
| Excel Modeling Benchmark v1.1 | 69% | #24 of 42 | Official leaderboard | |
| Vals Finance Agent v2.1 | 61.4% | #23 of 40 | Official leaderboard | |
| Vals Finance Index | 56.9% | #24 of 40 | Official leaderboard | |
| MedCode | 40.7% | #58 of 96 | Official leaderboard | |
| Excel Modeling Benchmark | 56.2% | #47 of 76 | Official leaderboard | |
| Vals Index v2.1 | 48.7% | #34 of 49 | Official leaderboard | |
| Immigration Bench | 63.7% | #19 of 25 | Official leaderboard | |
| Vals Legal Index | 65.6% | #31 of 38 | Official leaderboard |
Vision, audio and robotics
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Furniture Assembly Benchmark | 20.8% | #38 of 39 | Epoch AI run |
Safety and security
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| SAFE-Teen Relationship boundaries | 19% | #2 of 17 | Official leaderboard | |
| SAFE-Teen Urgent safety | 33% | #8 of 17 | Official leaderboard | |
| SAFE-Teen Health advice | 53% | #11 of 17 | Official leaderboard |
Mistral AI's models over time
Mistral AI's models in color, against every other model on the Epoch Capabilities Index, by release date.
Source: Epoch Capabilities Index, Epoch AI (CC BY 4.0)