EuroEval Romanian NLU - RONEC: leaderboard
Metric: Named entity recognition Score (%). Source: euroeval.com. 207 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Mistral Small 3.1 | 52.34 |
| 2 | Mistral Small 3.2 | 50.91 |
| 3 | Gemma 3 12B (IT) | 48.7 |
| 4 | Gemma 4 31B (IT) | 48.5 |
| 5 | Ministral 3 14B | 47.5 |
| 6 | Llama 3.3 70B Instruct | 47.17 |
| 7 | Gemini 2.5 Flash (Thinking) | 47.06 |
| 8 | Gemma 3 27B (IT) | 46.68 |
| 9 | GPT-5 | 46.17 |
| 10 | Qwen 3 4B 2507 (Thinking) | 45.92 |
| 11 | gemma-3-27B-pt | 45.91 |
| 12 | Olmo-3.1-32B-Instruct-SFT | 45.89 |
| 13 | Claude Sonnet 4.5 (Thinking) | 45.63 |
| 14 | GPT-OSS-20B (Low) | 45.55 |
| 15 | Ministral-3-14B-Reasoning-2512 | 45.42 |
Interactive version: theaggregate.ai/benchmark?slug=euroeval-romanian-nlu-ronec · How It Works · Data refreshed daily, snapshot 2026-09-05.