MEGA-Bench - Skill: Commonsense And Social Reasoning: leaderboard
Metric: Average Score (%). Source: huggingface.co. 44 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Pro (Preview 03-25) | 68.7 |
| 2 | GPT-4o | 63.5 |
| 3 | Claude 3.5 Sonnet (20241022) | 59.1 |
| 4 | InternVL3-78B | 58 |
| 5 | Claude 3.5 Sonnet (20240620) | 57.6 |
| 6 | Gemini 1.5 Pro (002) | 57.5 |
| 7 | Qwen 2 VL 72B | 56.7 |
| 8 | InternVL3-38B | 56.2 |
| 9 | Gemini 1.5 Flash (002) | 56.2 |
| 10 | InternVL2.5-78B | 56.1 |
| 11 | GPT-4o Mini | 55.7 |
| 12 | Gemma 3 27B (IT) | 54.3 |
| 13 | Gemma 3 12B (IT) | 52.1 |
| 14 | InternVL3-8B | 48.6 |
| 15 | Qwen 2 VL 7B | 46.3 |
Interactive version: theaggregate.ai/benchmark?slug=mega-bench-skill-commonsense-and-social-reasoning · How It Works · Data refreshed daily, snapshot 2026-09-05.