Mistral Large 4
Mistral Large 4 (Mistral AI) was released on Oct 6, 2026. In the Epoch AI capability ranking (ECI), it ranks #72 out of 267 models, with 146.90. On Arena, which measures human preference in blind comparisons, it has 1,494 ELO points (#18). On LiveBench, the overall average is 71.8, #45 out of 58 scored models. Against models of the same generation, its best result is on AIME 2024 and 2025 (OTIS): 98.3%, #22 of 75, tied with 3 others.
Where Mistral Large 4 stands out and where it falls behind, by the numbers
Mistral Large 4 has published results on 14 tests compiled by Epoch AI. The yardstick here is its generation: its position among the models released between Jan 9, 2026 and Jan 4, 2027 that took the same test. Against the whole archive, which includes 2023 models, a recent release lands near the top of almost everything, and that says nothing.
Where it stands out
Tests where it sits in the top third of its own generation.
Where it falls behind
Tests where it sits in the bottom half of its own generation.
See all 14 tests, one by one
| Test | Score | In its generation | Overall | Test record |
|---|---|---|---|---|
| GPQA Diamond | 88.1% | #49 / 73 | #52 / 173 | 96.4% Gemini 4 Argon |
| AIME 2024 and 2025 (OTIS) | 98.3% | #22 / 75 (tied with 3) | #22 / 168 | 100.0% GPT-5.5 |
| CritPt | 10.6% | #47 / 76 (tied with 1) | #48 / 112 | 32.3% GPT-5.6 Sol |
| SciCode | 54.2% | #32 / 74 | #32 / 104 | 66.9% Claude Opus 5.5 |
| Frontiermath Tiers 1 3 V2 | 58.6% | #39 / 64 | #41 / 86 | 93.7% GPT-6 Astra |
| ALE-Bench | 1,213 | #24 / 57 | #26 / 82 | 2,951 GPT-6 Astra |
| SimpleQA Verified | 20.0% | #55 / 59 | #69 / 81 | 76.0% Gemini 4 Argon |
| Chess puzzles | 33.0% | #36 / 58 (tied with 1) | #43 / 80 | 76.0% GPT-6 Astra |
| Game puzzles | 17.0% | #44 / 58 (tied with 3) | #51 / 77 | 84.0% GPT-6 Astra |
| ProofBench | 10.0% | #56 / 64 | #63 / 74 | 100.0% Claude Fable 5.1 |
| Frontiermath Tier 4 V2 | 26.8% | #32 / 54 (tied with 5) | #34 / 67 | 100.0% GPT-6.1 Sol |
| FrontierCode | 39.4% | #27 / 43 | #27 / 43 | 54.6% Claude Opus 5.5 |
| Furniture Assembly | 20.8% | #30 / 32 (tied with 1) | #32 / 34 | 83.3% Claude Opus 5.5 |
| Surface Evolver Bench | 28.1% | #22 / 23 | #22 / 24 | 95.0% Claude Fable 5 |
For each test, the model's best effort level counts, as in the ranking. Source of the scores: Epoch AI, under CC BY 4.0.
What to compare Mistral Large 4 with
The models from other labs closest in capability and released in the same period, at most two per lab. Click any of them to see its page.
| Model | Lab | ECI | ELO | LiveBench | Release |
|---|---|---|---|---|---|
| Qwen3.7-Plus | Alibaba | 147.30 | 1,510 | n/a | Jun 2, 2026 |
| MiniMax-M3 | MiniMax | 147.13 | 1,471 | 67.3 | Jun 1, 2026 |
| Mistral Large 4 | Mistral AI | 146.90 | 1,494 | 71.8 | Oct 6, 2026 |
| MiniMax-M2.5 | MiniMax | 146.67 | 1,393 | n/a | Feb 12, 2026 |
| Qwen3.5 397B-A17B | Alibaba | 146.66 | 1,499 | n/a | Feb 13, 2026 |
| DeepSeek v4 (unknown) | DeepSeek | 146.26 | n/a | n/a | Apr 24, 2026 |
| GPT-5.4 Nano | OpenAI | 146.12 | 1,404 | 69.6 | Mar 17, 2026 |
Mistral Large 4 or MiniMax-M3?
MiniMax-M3 is the model from another lab closest to Mistral Large 4 in capability. Across the 10 tests both took, this is the score:
| Test | Mistral Large 4 | MiniMax-M3 |
|---|---|---|
| CritPt | 10.6% | 3.7% |
| FrontierCode | 39.4% | 14.7% |
| Chess puzzles | 33.0% | 13.0% |
| Game puzzles | 17.0% | 8.0% |
| Surface Evolver Bench | 28.1% | 55.0% |
| ALE-Bench | 1,213 | 640 |
| ProofBench | 10.0% | 18.0% |
| AIME 2024 and 2025 (OTIS) | 98.3% | 71.1% |
| SciCode | 54.2% | 47.1% |
| GPQA Diamond | 88.1% | 90.9% |
Unweighted score: each test counts one point, whether math or coding. The dot marks who did better on each row.
Other models from Mistral AI
The models from the same lab closest in time, newest first.
| Model | Release | ECI | ELO | LiveBench | Access |
|---|---|---|---|---|---|
| Mistral Large 4 | Oct 6, 2026 | 146.90 | 1,494 | 71.8 | Unknown |
| Mistral Medium 3.5 | Apr 28, 2026 | 141.47 | 1,461 | n/a | Open weights (restricted use) |
| Magistral Small 1.2 | Sep 18, 2025 | 131.43 | n/a | n/a | Open weights (unrestricted) |
| Mistral Medium 3.1 | Aug 12, 2025 | 135.67 | n/a | n/a | API access |
| Mistral Small 3.2 | Jun 20, 2025 | 131.75 | n/a | n/a | Open weights (unrestricted) |
| Magistral Small 1.0 | Jun 10, 2025 | 133.20 | n/a | n/a | Open weights (unrestricted) |
| Mistral Medium 3 | May 7, 2025 | 134.06 | n/a | n/a | API access |
| Mistral Small 3.1 | Mar 17, 2025 | 127.49 | n/a | n/a | Open weights (unrestricted) |
| Mistral Small 3 | Jan 30, 2025 | 127.08 | n/a | n/a | Open weights (unrestricted) |
Have you used Mistral Large 4? Rate it
Users' opinion, not measurement: the rating is kept apart from the ranking and does not move anyone's position. One vote per browser, no sign-up, and you can change your rating later.
What people ask about Mistral Large 4
Where does Mistral Large 4 rank among AI models?
In the ROO3 AI Benchmark, ordered by the Epoch AI ECI, Mistral Large 4 ranks #72 out of 267 models, with an ECI of 146.90. The ECI is a composite index that combines several hard benchmarks into a single scale. The position changes when a new model comes out, and the page is refreshed six times a day.
Is Mistral Large 4 better than MiniMax-M3?
Across the 10 tests both models took and Epoch AI compiled, Mistral Large 4 comes out ahead on 7, behind on 3 and tied on 0. It is a test-by-test count with no weighting: a win in math counts the same as a win in coding. To choose between the two, look at the tests that resemble your own use.
What is Mistral Large 4 best at?
Against models of the same generation (released up to 9 months before and up to 3 months after it), the best relative results of Mistral Large 4 are: AIME 2024 and 2025 (OTIS) (#22 of 75). The yardstick is its position among the models of that period that took the same test.
Where do the numbers on this page come from?
From the same three sources as the ranking, all openly licensed: Epoch AI (ECI and individual tests, CC BY 4.0), Arena (human-preference ELO, CC BY 4.0) and LiveBench (overall average, CC BY-SA 4.0). ROO3 does not measure any model. When the page shows a figure published by the maker itself, it appears in the written analysis, labeled as the maker's, and never enters the ranking.
Where each number comes from
Epoch AI, "AI Benchmarking Hub". Published at epoch.ai. Data under the CC BY 4.0 licence, reordered by roo3.co.
The Arena (LMArena) leaderboard, from the public lmarena-ai/leaderboard-dataset on Hugging Face, under the CC BY 4.0 licence. We are not affiliated with the Arena.
LiveBench, by the LiveBench team, published at livebench.ai under the CC BY-SA 4.0 license. The overall average is recomputed by roo3.co with the same formula as the website (average of the categories), and those numbers stay under CC BY-SA 4.0. We are not affiliated with LiveBench.
← See the full ranking