EuroEval Ukrainian NLU - ScaLA UK: leaderboard
The "ScaLA UK" subtask of EuroEval Ukrainian NLU. Ukrainian-language EuroEval natural-language-understanding suite, separating NLU task performance from the broader all-task EuroEval aggregate.
Metric: Matthews correlation coefficient (%). Source: euroeval.com. Saturation forecast: Around February 2028. 201 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.1 Flash Lite (Preview) | 52.39 | #194 |
| 2 | Gemini 2.5 Flash (Thinking) | 46.76 | #245 (Gemini 2.5 Flash) |
| 3 | GPT-5.2 | 45.83 | #110 |
| 4 | Gemini 3 Pro (Preview) | 45.3 | #66 |
| 5 | Mistral Small 3.1 | 45.24 | #610 |
| 6 | Gemini 2.5 Flash Lite (Thinking) | 45.21 | #422 (Gemini 2.5 Flash Lite) |
| 7 | Qwen 2.5 72B Instruct | 44.58 | #454 |
| 8 | Ministral-3-14B-Instruct-2512 | 43.85 | #615 |
| 9 | Claude Sonnet 4.6 | 43.79 | #90 |
| 10 | Mistral Small 3.2 | 43.24 | #605 |
| 11 | Gemini 2.5 Flash Lite (Non-reasoning) | 42.26 | #422 (Gemini 2.5 Flash Lite) |
| 12 | Gemini 2.5 Flash (Non-reasoning) | 42.24 | #245 (Gemini 2.5 Flash) |
| 13 | MamayLM-Gemma-3-12B-IT-v1.0 | 42.16 | #584 |
| 14 | Ministral-3-14B-Reasoning-2512 | 41.67 | #540 |
| 15 | Llama 3.3 70B Instruct | 41.33 | #534 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=euroeval-ukrainian-nlu-scala-uk · How It Works · Data refreshed daily, snapshot 2026-10-20.