Introducing Mistral 3 Dec 2, 2025
Mistral Large 3
Mistral Large 3.
Mistral AI Released December 2nd, 2025 Family Mistral Large The lab's model card Compare with others 69
Benchmarks with a score
0
Ranked first, out of five or more models
0
In the top three
Mistral AI Dec 2025 Mistral Large 3
Open weights Flagship
No Epoch Capabilities Index score yet
- Chartography 9.00% #43 /43
- Vals Finance Agent 18.1% #45 /49
- CorpFin v2 61% #61 /119
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 | |
|---|---|---|---|---|
| HUMAINE | 33 | #6 of 50 | Official leaderboard | |
| LMArena (Elo) | 1,415 | #113 of 335 | Official leaderboard | |
| CulturalRiddles | 38.5% | #7 of 20 | Official leaderboard | |
| HELM Arabic | 79.0 | #12 of 28 | Official leaderboard | |
| HELM Arabic Enterprise | 66.0 | #10 of 23 | Official leaderboard | |
| MMMLU | 85.5% | #31 of 71 | Lab model card | |
| SimpleQA | 23.8% | #34 of 60 | Lab model card | |
| Judgemark v2.1 | 68.2 | #24 of 42 | Official leaderboard | |
| Creative Writing v3 | 62.3 | #77 of 120 | Official leaderboard | |
| Longform Creative Writing (EQ-Bench) | 48.3 | #87 of 119 | Official leaderboard | |
| Lech Mazur Creative Writing | -1.81 | #38 of 50 | Official leaderboard | |
| Phare | 60.9% | #48 of 62 | Official leaderboard | |
| Antidote | 947 | #31 of 36 | Official leaderboard | |
| Vectara Hallucination Leaderboard (2025 dataset) | 14.5% | #92 of 104 | Official leaderboard | |
| DesignArena ASCII Art | 1,064 | #72 of 77 | Official leaderboard | |
| Extended NYT Connections | 7.50% | #101 of 103 | Official leaderboard | |
| BullshitBench V2 | 2.00% | #129 of 129 | Official leaderboard | |
| BullshitBench | 3.64% | #107 of 107 | Official leaderboard |
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| LiveMathematicianBench | 28.3% | #8 of 17 | Official leaderboard | |
| AMC | 52% | #2 of 4 | Lab model card | |
| DTBench | 65.1% | #125 of 180 | Official leaderboard | |
| GPQA Diamond | 43.9% | #261 of 302 | Lab model card | |
| Verifying the Verifiers (Project APE CRED benchmark) | 32% | #20 of 23 | Official leaderboard | |
| SimpleBench | 20.4% | #98 of 106 | Official leaderboard |
Coding
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| DesignArena Website | 1,155 | #101 of 151 | Official leaderboard | |
| DesignArena Code | 1,132 | #103 of 143 | Official leaderboard | |
| DesignArena 3D Design | 1,082 | #99 of 136 | Official leaderboard | |
| DesignArena UI Component | 1,093 | #103 of 139 | Official leaderboard | |
| DesignArena Game Dev | 1,070 | #110 of 143 | Official leaderboard | |
| DesignArena Data Visualization | 1,134 | #96 of 118 | Official leaderboard | |
| IOI (Vals) | 4.00% | #45 of 55 | Official leaderboard | |
| Code Arena | 1,233 | #120 of 128 | Official leaderboard | |
| DesignArena SVG | 983 | #97 of 103 | Official leaderboard | |
| WebDev Arena | 1,230 | #99 of 105 | arena.ai, via Epoch AI | |
| Image-to-WebDev Arena | 1,362 | #53 of 55 | Official leaderboard | |
| Snorkel Agentic Coding | 13.8% | #11 of 11 | Official leaderboard |
Agents and tools
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Elimination Game | 4 | #33 of 56 | Official leaderboard | |
| PinchBench | 65.9% | #31 of 48 | Official leaderboard | |
| Step Race | 1 | #41 of 59 | Official leaderboard | |
| Pencil Puzzle Bench | 0% | #41 of 53 | Official leaderboard | |
| PinchBench 2.0 | 54.5% | #46 of 57 | Official leaderboard | |
| clembench | 43.1 | #5 of 6 | Official leaderboard | |
| PACT | 1,449 | #24 of 26 | Official leaderboard |
Professional work
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| CaseLaw v2 | 61.4% | #21 of 53 | Official leaderboard | |
| TaxEval v2 | 73.1% | #52 of 129 | Official leaderboard | |
| CorpFin v2 | 61% | #61 of 119 | Official leaderboard | |
| Vals Finance Agent v1 | 23.5% | #31 of 47 | Official leaderboard | |
| LegalBench | 79.1% | #93 of 138 | Official leaderboard | |
| MortgageTax | 52.1% | #75 of 87 | Official leaderboard | |
| DAYJOB: Finance | 0% | #35 of 39 | Official leaderboard | |
| Vals Finance Agent | 18.1% | #45 of 49 | Official leaderboard | |
| DAYJOB: Healthcare | 0% | #34 of 37 | Official leaderboard | |
| EnterpriseBench: CoreCraft | 3.60% | #51 of 51 | Official leaderboard | |
| Tuesday Work Index | 13.5 | #44 of 44 | Official leaderboard | |
| GDP.xlsx | 0% | #20 of 20 | Official leaderboard | |
| RadLE 2.0 | 181.0 | #12 of 12 | Official leaderboard |
Vision, audio and robotics
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Vision Arena | 1,200 | #79 of 136 | Official leaderboard | |
| Chartography | 9.00% | #43 of 43 | Official leaderboard |
Safety and security
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| BioTIER Permit | 99.9% | #5 of 69 | Official leaderboard | |
| LiveSecBench | 28.4 | #28 of 34 | Official leaderboard | |
| Human Pathogen Capabilities Test v2 (HPCT v2) | 30.8% | #64 of 71 | Official leaderboard | |
| DistressBench | 58.8% | #40 of 42 | Official leaderboard | |
| Lakera B3 (Backbone Breaker Benchmark) | 39.3% | #34 of 35 | Official leaderboard | |
| KORA | 10.8% | #41 of 42 | Official leaderboard | |
| BioTIER Refuse | 9.20% | #69 of 69 | Official leaderboard |
Other
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| ACCoRD | 48.7 | #115 of 152 | Official leaderboard | |
| Conceptual Reasoning Index | 28.1 | #120 of 151 | Official leaderboard | |
| LMCA | 16.7% | #121 of 147 | Official leaderboard | |
| SAGE (Vals) | 24.6% | #80 of 81 | Official leaderboard |
Its lab's number next to a leaderboard
Where the benchmark's own leaderboard, or Epoch AI's run, also has this model. Gaps mostly come from different harnesses, effort levels and task subsets. Every pair.
| Benchmark | Lab says | Board says | Gap | What differs |
|---|---|---|---|---|
| LMArena (Elo) | 1,418 LMArena ELO Score (+/- 11) | 1,415 Leaderboard | +3 | Mistral quotes its December 2025 Arena rating, while the board reflects a later rating snapshot. |
Its launch post
The benchmarks the lab chose to headline, with the scores it gave.