EuroEval Romanian Knowledge — leaderboard
Metric: Knowledge Average Score (%). Source: euroeval.com. 190 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 91.92 |
| 2 | GPT-5 | 90.62 |
| 3 | Claude Sonnet 4.5 (Thinking) | 90.36 |
| 4 | GPT-5.4 Mini (Medium) | 89.98 |
| 5 | Gemini 2.5 Pro | 89.79 |
| 6 | Gemini 3 Flash (Preview) (Non-reasoning) | 89.53 |
| 7 | Grok 4.1 Fast (Reasoning) | 88.81 |
| 8 | GPT-5.4 Mini (High) | 88.06 |
| 9 | GPT-5.4 Nano (High) | 87.97 |
| 10 | GPT-5 Nano | 86.8 |
| 11 | GPT-5 Mini | 85.76 |
| 12 | Gemini 2.5 Flash (Thinking) | 85.75 |
| 13 | GPT-5.4 Nano (Medium) | 82.94 |
| 14 | Claude Sonnet 4.6 | 82.77 |
| 15 | Qwen 3.5 9B | 81.31 |
Interactive version: theaggregate.ai/benchmark?slug=euroeval-romanian-knowledge · How the rankings work · Data refreshed daily, snapshot 2026-07-22.