EuroEval Spanish NLU - MLQA ES: leaderboard
The "MLQA ES" subtask of EuroEval Spanish NLU. Spanish-language EuroEval natural-language-understanding suite, separating NLU task performance from the broader all-task EuroEval aggregate.
Metric: Character F1 (%). Source: euroeval.com. Saturation forecast: Not forecast. 411 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Bielik-11B-v2.3-Instruct | 72.14 | #693 |
| 2 | Llama 3.1 405B | 70.79 | #466 |
| 3 | SOLAR-10.7B-v1.0 | 70.14 | #1253 |
| 4 | Yi 1.5 34B | 69.63 | #807 |
| 5 | Ministral-3-14B-Reasoning-2512 | 69.56 | #540 |
| 6 | Magistral Small 1.2 | 69.43 | #570 |
| 7 | Qwen3.5-4B-Base | 69.36 | #798 |
| 8 | Gemma 2 27B | 69.34 | #806 |
| 9 | OLMo-2-1124-7B | 69.06 | #1401 |
| 10 | Llama 3 70B | 68.93 | #631 |
| 11 | Llama 2 70B Base | 68.79 | #1025 |
| 12 | Qwen 3 32B (Non-reasoning) | 68.67 | #443 (Qwen 3 32B) |
| 13 | Mistral Small 3.2 | 68.58 | #605 |
| 14 | Llama 3.1 70B | 68.51 | #592 |
| 15 | Qwen 3 14B (Non-reasoning) | 68.1 | #537 (Qwen 3 14B) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=euroeval-spanish-nlu-mlqa-es · How It Works · Data refreshed daily, snapshot 2026-10-20.