EuroEval French NLU - ScaLA FR — leaderboard
Metric: Linguistic acceptability Score (%). Source: euroeval.com. 377 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 75.59 |
| 2 | Gemma 4 31B (IT) | 73.85 |
| 3 | Ministral-3-14B-Reasoning-2512 | 73.6 |
| 4 | Gemini 3.1 Flash Lite (Preview) | 72.61 |
| 5 | Gemini 3 Flash (Preview) (Non-reasoning) | 72.36 |
| 6 | Claude Sonnet 4.6 | 72.12 |
| 7 | Ministral 3 14B | 70.9 |
| 8 | Ministral-3-14B-Instruct-2512 | 70.67 |
| 9 | Gemma 4 26B A4B (IT) | 70.52 |
| 10 | Ministral-3-8B-Instruct-2512 | 68.28 |
| 11 | GPT-5.2 | 67.46 |
| 12 | Llama 3.1 70B | 66.59 |
| 13 | Ministral 3 8B | 66.43 |
| 14 | GPT-5.4 Mini (Non-reasoning) | 65.47 |
| 15 | GPT-5.4 Mini (Medium) | 63.5 |
Interactive version: theaggregate.ai/benchmark?slug=euroeval-french-nlu-scala-fr · How the rankings work · Data refreshed daily, snapshot 2026-07-22.