EuroEval Latvian NLU - ScaLA LV: leaderboard
Metric: Linguistic acceptability Score (%). Source: euroeval.com. 233 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 58.1 |
| 2 | GPT-5.2 | 56.85 |
| 3 | Gemini 2.5 Flash (Thinking) | 56.6 |
| 4 | Gemini 3.1 Flash Lite (Preview) | 52.6 |
| 5 | GPT-5 | 50.23 |
| 6 | Claude Sonnet 4.6 | 49.49 |
| 7 | Gemini 2.5 Pro | 49.04 |
| 8 | GPT-4.1 | 46.08 |
| 9 | Qwen 3 235B A22B 2507 Instruct | 44.64 |
| 10 | GPT-5.4 Mini (Medium) | 44.39 |
| 11 | GPT-5.4 Nano (High) | 42.54 |
| 12 | GPT-5.4 Nano (Medium) | 41.91 |
| 13 | GPT-5 Mini | 41.23 |
| 14 | Claude Sonnet 4.5 (Thinking) | 40.63 |
| 15 | GPT-5.4 Mini (High) | 40.49 |
Interactive version: theaggregate.ai/benchmark?slug=euroeval-latvian-nlu-scala-lv · How It Works · Data refreshed daily, snapshot 2026-09-05.