EuroEval Polish NLU - ScaLA PL: leaderboard
Metric: Linguistic acceptability Score (%). Source: euroeval.com. 234 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 71.7 |
| 2 | Gemini 3.1 Flash Lite (Preview) | 67.21 |
| 3 | Gemini 2.5 Flash (Thinking) | 65.39 |
| 4 | GPT-5.4 Mini (Medium) | 64.15 |
| 5 | Claude Sonnet 4.6 | 61.86 |
| 6 | GPT-5.2 | 61.36 |
| 7 | GPT-5.4 Mini (High) | 60.74 |
| 8 | GPT-5 | 59.8 |
| 9 | Gemma 4 31B (IT) | 59.77 |
| 10 | Gemini 2.5 Pro | 59.32 |
| 11 | Claude Sonnet 4.5 (Thinking) | 56.9 |
| 12 | GPT-5.4 Nano (High) | 55.49 |
| 13 | Gemma 4 26B A4B (IT) | 55.07 |
| 14 | GPT-5 Nano | 54.8 |
| 15 | Grok 4.1 Fast (Reasoning) | 54.51 |
Interactive version: theaggregate.ai/benchmark?slug=euroeval-polish-nlu-scala-pl · How It Works · Data refreshed daily, snapshot 2026-09-05.