EuroEval Greek NLU - Greek SA: leaderboard
The "Greek SA" subtask of EuroEval Greek NLU. Greek-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: Estimated already saturated. 204 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | gemma-3-12B-pt | 84 | #686 |
| 2 | Llama 3.3 70B Instruct | 83.74 | #534 |
| 3 | Llama 3.1 70B | 83.3 | #592 |
| 4 | Gemma 3 27B (IT) | 82.43 | #523 |
| 5 | Gemini 3 Flash (Preview) (Thinking) | 81.97 | #82 (Gemini 3 Flash (Preview)) |
| 6 | Gemma 4 31B (IT) | 81.64 | #176 |
| 7 | gemma-3-27B-pt | 81.08 | #502 |
| 8 | Qwen 3 30B A3B 2507 Instruct | 80.73 | #488 |
| 9 | Grok 4.1 Fast (Reasoning) | 80.68 | #217 (Grok 4.1 Fast) |
| 10 | OLMo-2-0325-32B-Instruct | 80.66 | #848 |
| 11 | Gemma 3 12B (IT) | 80.58 | #664 |
| 12 | GPT-4.1 | 80.5 | #248 |
| 13 | Qwen 3 32B (Non-reasoning) | 80.07 | #443 (Qwen 3 32B) |
| 14 | Claude Sonnet 4.5 (Thinking) | 79.99 | #142 (Claude Sonnet 4.5) |
| 15 | Claude Sonnet 4.6 | 79.79 | #90 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=euroeval-greek-nlu-greek-sa · How It Works · Data refreshed daily, snapshot 2026-10-20.