LA Leaderboard - Belebele Catalan: leaderboard
Belebele four-option reading comprehension on FLORES-200 passages, Catalan version. From the LA Leaderboard, which evaluates models on Spanish, Catalan, Basque and Galician tasks with the lm-evaluation-harness; accuracy (%), higher is better. Each model result file is a separate run.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 69 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemma 2 9B (IT) | 90.33 | #795 |
| 2 | Gemma 2 9B | 86.89 | #924 |
| 3 | Llama 3.1 8B Instruct | 86 | #1017 |
| 4 | Qwen 2.5 7B | 85.56 | #895 |
| 5 | Qwen 2.5 7B Instruct | 84.33 | #863 |
| 6 | Yi-1.5-9B | 82 | #1215 |
| 7 | Qwen 2.5 3B Instruct | 80.89 | #1159 |
| 8 | Yi-1.5-9B Chat | 80.67 | #1032 |
| 9 | Qwen 2.5 3B | 80.22 | #1177 |
| 10 | Llama 3.1 8B | 79.44 | #1130 |
| 11 | Gemma 3 4B (IT) | 78.33 | #997 |
| 12 | aya-expanse-8B | 78.22 | #914 |
| 13 | Mistral 7B Instruct (v0.3) | 76.33 | #1255 |
| 14 | Mistral-7B-v0.3 | 75.67 | #1321 |
| 15 | EuroLLM-9B | 74.78 | #1238 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=la-leaderboard-belebele-catalan · How It Works · Data refreshed daily, snapshot 2026-11-05.