LA Leaderboard - XQuAD Catalan: leaderboard
XQuAD extractive question answering in Catalan. From the LA Leaderboard, which evaluates models on Spanish, Catalan, Basque and Galician tasks with the lm-evaluation-harness; F1 (%), higher is better. Each model result file is a separate run.
Metric: F1 (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 69 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 2.5 7B | 80.03 | #895 |
| 2 | Yi-1.5-9B | 77.5 | #1215 |
| 3 | EuroLLM-9B | 75.09 | #1238 |
| 4 | Llama 3.1 8B | 74.64 | #1130 |
| 5 | Gemma 2 9B | 74.46 | #924 |
| 6 | OLMo-2-1124-7B | 73.66 | #1381 |
| 7 | Qwen 2.5 3B | 71.76 | #1177 |
| 8 | Mistral-7B-v0.3 | 71.7 | #1321 |
| 9 | aya-expanse-8B | 71.6 | #914 |
| 10 | gemma-3-4B-pt Base | 70.79 | #1110 |
| 11 | occiglot-7B-eu5 | 70.5 | |
| 12 | Gemma 3 4B (IT) | 69.37 | #997 |
| 13 | salamandra-7B | 68.33 | #1446 |
| 14 | Llama 3.1 8B Instruct | 67.37 | #1017 |
| 15 | Llama 3.2 3B | 67.09 | #1376 |
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-xquad-catalan · How It Works · Data refreshed daily, snapshot 2026-11-05.