EuroEval Catalan NLU - ScaLA CA: leaderboard
The "ScaLA CA" subtask of EuroEval Catalan NLU. Catalan-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 March 2027. 212 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 71.6 | #66 |
| 2 | GPT-5.2 | 71.59 | #110 |
| 3 | Claude Sonnet 4.6 | 68.77 | #90 |
| 4 | Gemini 3 Flash (Preview) (Non-reasoning) | 66.37 | #82 (Gemini 3 Flash (Preview)) |
| 5 | Gemini 3.1 Flash Lite (Preview) | 64.19 | #194 |
| 6 | Qwen 3 32B (Non-reasoning) | 63.84 | #443 (Qwen 3 32B) |
| 7 | Gemini 2.5 Flash (Thinking) | 63.33 | #245 (Gemini 2.5 Flash) |
| 8 | GPT-4.1 | 62.88 | #248 |
| 9 | Gemini 3 Flash (Preview) (Thinking) | 62.29 | #82 (Gemini 3 Flash (Preview)) |
| 10 | Gemini 2.5 Pro | 61.38 | #152 |
| 11 | Qwen 2.5 72B Instruct | 58.66 | #454 |
| 12 | Gemini 2.5 Flash Lite (Thinking) | 57.82 | #422 (Gemini 2.5 Flash Lite) |
| 13 | Qwen 3.5 35B A3B Base | 55.52 | #310 |
| 14 | Qwen 3 30B A3B 2507 Instruct | 54.92 | #488 |
| 15 | GPT-4.1 Mini | 54.78 | #355 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=euroeval-catalan-nlu-scala-ca · How It Works · Data refreshed daily, snapshot 2026-10-20.