EuroEval Swedish Summarization - Swedn: leaderboard
EuroEval's Swedish summarization task on the Swedn dataset — models write abstractive summaries of Swedish news articles, scored with automatic text-overlap metrics, from the EuroEval European language model benchmark suite.
Metric: chrF3++ (%). Source: euroeval.com. Saturation forecast: Around 2034. 153 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemma 4 31B (IT) | 39.77 | #176 |
| 2 | Gemini 3 Pro (Preview) | 39.05 | #66 |
| 3 | Gemma 4 26B A4B (IT) | 39.04 | #297 |
| 4 | GPT-4.1 | 38.91 | #248 |
| 5 | Bielik-11B-v2.3-Instruct | 38.8 | #693 |
| 6 | Apertus-70B-Instruct-2509 | 38.69 | #820 |
| 7 | EuroLLM-9B-Instruct | 38.53 | #1054 |
| 8 | Llama 3.3 70B Instruct | 38.48 | #534 |
| 9 | Claude Sonnet 4.5 (Non-reasoning) | 38.4 | #142 (Claude Sonnet 4.5) |
| 10 | Qwen 3.5 9B | 38.24 | #379 |
| 11 | Gemini 3 Flash (Preview) (Non-reasoning) | 38.2 | #82 (Gemini 3 Flash (Preview)) |
| 12 | Ministral-3-14B-Instruct-2512 | 38.13 | #615 |
| 13 | Mistral Small 3.1 | 37.97 | #610 |
| 14 | EuroLLM-22B-Instruct-2512 | 37.83 | #701 |
| 15 | Claude Sonnet 4.5 (Thinking) | 37.81 | #142 (Claude Sonnet 4.5) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=euroeval-swedish-summarization-swedn · How It Works · Data refreshed daily, snapshot 2026-10-20.