EuroEval Spanish Summarization - Mlsum ES: leaderboard
EuroEval's Spanish summarization task on the Mlsum ES dataset — models write abstractive summaries of Spanish news articles, scored with automatic text-overlap metrics, from the EuroEval European language model benchmark suite.
Metric: chrF3++ (%). Source: euroeval.com. Saturation forecast: Not forecast. 148 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | occiglot-7B-eu5-instruct | 31.28 | |
| 2 | Bielik-11B-v2.3-Instruct | 31.19 | #693 |
| 3 | Gemma 3 27B (IT) | 29.68 | #523 |
| 4 | Qwen 2.5 72B Instruct | 29.62 | #454 |
| 5 | cogito-v1-preview-qwen-14B | 29.55 | |
| 6 | Apertus-70B-Instruct-2509 | 29.34 | #820 |
| 7 | granite-4.0-h-micro | 29.33 | #1088 |
| 8 | Qwen 3 8B (Non-reasoning) | 29.13 | #674 (Qwen 3 8B) |
| 9 | Qwen 3 14B (Non-reasoning) | 29.12 | #537 (Qwen 3 14B) |
| 10 | tiny-aya-water | 29.02 | |
| 11 | Tiny Aya Global | 28.88 | #1325 |
| 12 | Llama 3.1 8B Instruct | 28.87 | #1035 |
| 13 | Llama 3.3 70B Instruct | 28.82 | #534 |
| 14 | Claude Sonnet 4.5 (Non-reasoning) | 28.69 | #142 (Claude Sonnet 4.5) |
| 15 | Qwen 3 30B A3B 2507 Instruct | 28.65 | #488 |
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-summarization-mlsum-es · How It Works · Data refreshed daily, snapshot 2026-10-20.