EuroEval Albanian Summarization - LR SUM SQ: leaderboard
EuroEval's Albanian summarization task on the LR SUM SQ dataset — models write abstractive summaries of Albanian 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. 154 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Tiny Aya Global | 35.09 | #1325 |
| 2 | tiny-aya-water | 34.8 | |
| 3 | Mistral Small 3.1 | 34.73 | #610 |
| 4 | Llama 3.1 8B Instruct | 34.7 | #1035 |
| 5 | Claude Sonnet 4.5 (Non-reasoning) | 34.68 | #142 (Claude Sonnet 4.5) |
| 6 | Bielik-11B-v2.3-Instruct | 34.61 | #693 |
| 7 | Qwen 3 4B (Non-reasoning) | 34.22 | #829 (Qwen 3 4B) |
| 8 | Qwen 3.5 27B FP8 | 34.07 | #164 |
| 9 | Gemini 2.5 Flash Lite (Non-reasoning) | 34.05 | #422 (Gemini 2.5 Flash Lite) |
| 10 | Gemma 4 E2B (IT) | 33.87 | #781 |
| 11 | Apertus-70B-Instruct-2509 | 33.87 | #820 |
| 12 | Gemma 3 270M (IT) | 33.82 | #1581 |
| 13 | GPT-4.1 | 33.78 | #248 |
| 14 | Gemma 4 31B (IT) | 33.78 | #176 |
| 15 | gemma-4-E4B-it | 33.75 | #566 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=euroeval-albanian-summarization-lr-sum-sq · How It Works · Data refreshed daily, snapshot 2026-10-20.