EuroEval French NLU - ScaLA FR: leaderboard
The "ScaLA FR" subtask of EuroEval French NLU. French-language EuroEval natural-language-understanding suite, separating NLU task performance from the broader all-task EuroEval aggregate.
Metric: Matthews correlation coefficient (%). Source: euroeval.com. Saturation forecast: Around January 2027. 377 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 75.59 | #66 |
| 2 | Gemma 4 31B (IT) | 73.85 | #176 |
| 3 | Ministral-3-14B-Reasoning-2512 | 73.6 | #540 |
| 4 | Gemini 3.1 Flash Lite (Preview) | 72.61 | #194 |
| 5 | Gemini 3 Flash (Preview) (Non-reasoning) | 72.36 | #82 (Gemini 3 Flash (Preview)) |
| 6 | Claude Sonnet 4.6 | 72.12 | #90 |
| 7 | Gemini 3 Flash (Preview) (Thinking) | 71.28 | #82 (Gemini 3 Flash (Preview)) |
| 8 | Ministral-3-14B-Instruct-2512 | 70.67 | #615 |
| 9 | Gemma 4 26B A4B (IT) | 70.52 | #297 |
| 10 | Ministral-3-8B-Instruct-2512 | 68.28 | #661 |
| 11 | GPT-5.2 | 67.46 | #110 |
| 12 | Qwen 3.5 35B A3B Base | 67.36 | #310 |
| 13 | Llama 3.1 70B | 66.59 | #592 |
| 14 | NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 | 61.6 | #679 |
| 15 | gemma-3-27B-pt | 61.16 | #502 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=euroeval-french-nlu-scala-fr · How It Works · Data refreshed daily, snapshot 2026-10-20.