Greek MMLU - Social Sciences: leaderboard
The Social Sciences subtask of Greek MMLU — social-science questions in Greek; accuracy (%).
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 84 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Flash (Preview) | 94.16 | #82 |
| 2 | GPT-5.2 | 90.29 | #110 |
| 3 | GPT-4o (2024-11-20) | 89.36 | #381 |
| 4 | Llama 3.3 70B Instruct | 82.92 | #534 |
| 5 | Llama 3.1 70B Instruct | 82.57 | #561 |
| 6 | Gemma 3 27B (IT) | 82.19 | #523 |
| 7 | Qwen 2.5 72B Instruct | 81.92 | #454 |
| 8 | Qwen 2.5 72B | 81.35 | #499 |
| 9 | gemma-3-27B-pt | 80.52 | #502 |
| 10 | Qwen 3 30B A3B 2507 Instruct | 79.33 | #488 |
| 11 | Gemma 3 12B (IT) | 79.28 | #664 |
| 12 | Llama 3.1 70B | 78.9 | #592 |
| 13 | gemma-3-12B-pt | 78.28 | #686 |
| 14 | Qwen 2.5 32B Instruct | 76.61 | #498 |
| 15 | Qwen 2.5 32B | 75.26 | #546 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=greek-mmlu-social-sciences · How It Works · Data refreshed daily, snapshot 2026-10-20.