EuroEval Lithuanian Summarization - Lrytas: leaderboard
EuroEval's Lithuanian summarization task on the Lrytas dataset — models write abstractive summaries of Lithuanian news articles, scored with automatic text-overlap metrics, from the EuroEval European language model benchmark suite.
Metric: chrF3++ (%). Source: euroeval.com. Saturation forecast: Around 2035. 148 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 35.19 | #66 |
| 2 | GPT-4.1 | 34.83 | #248 |
| 3 | Apertus-70B-Instruct-2509 | 34.41 | #820 |
| 4 | Gemma 3 27B (IT) | 34.17 | #523 |
| 5 | Gemini 3 Flash (Preview) (Non-reasoning) | 34.13 | #82 (Gemini 3 Flash (Preview)) |
| 6 | Qwen 3 30B A3B 2507 Instruct | 33.95 | #488 |
| 7 | GPT-5.4 Mini (2026-03-17) (High) | 33.91 | #169 (GPT-5.4 Mini (2026-03-17)) |
| 8 | Qwen 3 32B (Non-reasoning) | 33.84 | #443 (Qwen 3 32B) |
| 9 | EuroLLM-9B-Instruct | 33.75 | #1054 |
| 10 | Gemma 3n E4B (IT) | 33.55 | #749 |
| 11 | Gemini 3.1 Flash Lite (Preview) | 33.51 | #194 |
| 12 | Gemma 3 12B (IT) | 33.45 | #664 |
| 13 | Qwen 3 14B (Non-reasoning) | 33.38 | #537 (Qwen 3 14B) |
| 14 | Gemini 3 Flash (Preview) (Thinking) | 33.33 | #82 (Gemini 3 Flash (Preview)) |
| 15 | Qwen 2.5 72B Instruct | 33.26 | #454 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=euroeval-lithuanian-summarization-lrytas · How It Works · Data refreshed daily, snapshot 2026-10-20.