France: AI model benchmarks
AI models developed in France. 100 models tracked, 95 with a unified ELO rating. Current best: Mistral-Medium-3.5-128B (ELO 1631, rank #177 overall). Most recent tracked release: 2026-04-29.
Rated models
| Rank | Model | Provider | Unified ELO | Benchmarks |
|---|---|---|---|---|
| 177 | Mistral-Medium-3.5-128B | Mistral | 1631 ± 1 | 42 |
| 197 | Mistral-Large-3-675B-Instruct-2512 | Mistral | 1622 ± 1 | 18 |
| 285 | Mixtral_34Bx2_MoE_60B | Mistral | 1592 ± 1 | 14 |
| 296 | Mistral Medium 3.5 | Mistral | 1588 ± 1 | 164 |
| 325 | Mistral Medium 3.1 | Mistral | 1577 ± 1 | 452 |
| 335 | Devstral 2 | Mistral | 1575 ± 1 | 62 |
| 375 | Magistral Medium 1.2 | Mistral | 1567 ± 1 | 80 |
| 388 | Mistral Large 3 | Mistral | 1563 ± 1 | 320 |
| 405 | Devstral Small 2 | Mistral | 1558 ± 1 | 103 |
| 408 | Mistral-Nemo-Instruct-2407-abliterated | Mistral | 1558 ± 1 | 11 |
| 425 | Mistral Small 4 | Mistral | 1551 ± 1 | 137 |
| 442 | Mistral Large 2 (Jul) | Mistral | 1548 ± 1 | 164 |
| 448 | Mistral Large 2 (Nov) Instruct (2411) | Mistral | 1547 ± 1 | 155 |
| 463 | zephyr-orpo-141B-A35B-v0.1 | HuggingFace | 1544 ± 1 | 18 |
| 477 | pixtral-large-2411 | Mistral | 1540 ± 1 | 80 |
Interactive version: theaggregate.ai/country?slug=france · How It Works · Data refreshed daily, snapshot 2026-09-05.