EuroEval Portuguese Summarization - Publico: leaderboard
Metric: Score (%). Source: euroeval.com. 150 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Apertus-70B-Instruct-2509 | 34.74 |
| 2 | Qwen 3 30B A3B 2507 Instruct | 34.58 |
| 3 | Qwen 2.5 72B Instruct | 34.58 |
| 4 | GPT-5.2 | 34.41 |
| 5 | GPT-4.1 | 34.38 |
| 6 | Ministral-3-8B-Instruct-2512 | 33.81 |
| 7 | Claude Sonnet 4.5 (Thinking) | 33.74 |
| 8 | Ministral-3-14B-Instruct-2512 | 33.56 |
| 9 | Claude Sonnet 4.6 | 33.44 |
| 10 | Claude Haiku 4.5 (20251001) | 33.44 |
| 11 | Claude Sonnet 4.5 (Non-reasoning) | 33.4 |
| 12 | GPT-5 Mini | 33.37 |
| 13 | Ministral-3-3B-Reasoning-2512 | 33.35 |
| 14 | Gemini 3.1 Flash Lite (Preview) | 33.18 |
| 15 | GPT-5.4 Nano (High) | 33.09 |
Interactive version: theaggregate.ai/benchmark?slug=euroeval-portuguese-summarization-publico · How It Works · Data refreshed daily, snapshot 2026-09-05.