LA Leaderboard - SIQA Catalan: leaderboard
Catalan translation of Social IQa: commonsense reasoning about people's actions and motives. 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 | Mistral 7B Instruct (v0.3) | 50.56 | #1255 |
| 2 | Gemma 2 9B (IT) | 49.9 | #795 |
| 3 | salamandra-7B-instruct | 49.54 | #1392 |
| 4 | Llama 3.1 8B Instruct | 49.49 | #1017 |
| 5 | EuroLLM-9B-Instruct | 48.62 | #1047 |
| 6 | salamandra-7B | 48.21 | #1446 |
| 7 | Gemma 3 4B (IT) | 47.49 | #997 |
| 8 | Gemma 2 9B | 47.13 | #924 |
| 9 | EuroLLM-9B | 47.13 | #1238 |
| 10 | Qwen3-8B-Base | 46.88 | #626 |
| 11 | Mistral-7B-v0.3 | 46.88 | #1321 |
| 12 | occiglot-7B-eu5-instruct | 45.6 | |
| 13 | Llama 3.1 8B | 45.24 | #1130 |
| 14 | gemma-3-4B-pt Base | 44.78 | #1110 |
| 15 | occiglot-7B-eu5 | 44.63 |
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-siqa-catalan · How It Works · Data refreshed daily, snapshot 2026-11-05.