LA Leaderboard - XNLI Basque: leaderboard
Natural language inference on premises and hypotheses written natively in Basque (XNLIeu native). 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 | 53.46 | #924 |
| 2 | Gemma 2 9B (IT) | 46.54 | #795 |
| 3 | salamandra-2B | 45.25 | #1520 |
| 4 | salamandra-7B-instruct | 44.77 | #1392 |
| 5 | Llama 3.1 8B | 44.44 | #1130 |
| 6 | salamandra-7B | 44.44 | #1446 |
| 7 | salamandra-2B-instruct | 42.03 | |
| 8 | Llama 3.1 8B Instruct | 41.71 | #1017 |
| 9 | Qwen3-8B-Base | 41.55 | #626 |
| 10 | Gemma 3 4B (IT) | 40.42 | #997 |
| 11 | gemma-3-4B-pt Base | 40.26 | #1110 |
| 12 | gemma-3-1B-pt Base | 39.45 | #1519 |
| 13 | Llama 3.2 3B | 38.16 | #1376 |
| 14 | Gemma 3 1B (IT) | 37.36 | #1481 |
| 15 | aya-expanse-8B | 37.36 | #914 |
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-xnli-basque · How It Works · Data refreshed daily, snapshot 2026-11-05.