EuroEval Romanian NLU - RoSent: leaderboard
Metric: Sentiment classification Score (%). Source: euroeval.com. 207 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5 Mini | 98.51 |
| 2 | Llama 3.3 70B Instruct | 98.47 |
| 3 | Gemma 3 27B (IT) | 97.74 |
| 4 | Qwen 2.5 72B Instruct | 97.62 |
| 5 | Claude Sonnet 4.5 (Non-reasoning) | 97.6 |
| 6 | Mistral Small 3.1 | 97.56 |
| 7 | Grok 4.1 Fast (Reasoning) | 97.5 |
| 8 | Gemma 3 12B (IT) | 97.15 |
| 9 | GPT-5 | 97.1 |
| 10 | Mistral Small 3.2 | 96.97 |
| 11 | Claude Sonnet 4.5 (Thinking) | 96.8 |
| 12 | Gemini 3 Pro (Preview) | 96.79 |
| 13 | Qwen 3 30B A3B 2507 Instruct | 96.77 |
| 14 | gemma-3-12B-pt | 96.64 |
| 15 | Gemma 4 26B A4B (IT) | 96.64 |
Interactive version: theaggregate.ai/benchmark?slug=euroeval-romanian-nlu-rosent · How It Works · Data refreshed daily, snapshot 2026-09-05.