MEGA-Bench - Output Format: Open Ended Output: leaderboard
Metric: Average Score (%). Source: huggingface.co. 44 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Pro (Preview 03-25) | 70.4 |
| 2 | GPT-4o | 62.7 |
| 3 | Claude 3.5 Sonnet (20241022) | 62.4 |
| 4 | Claude 3.5 Sonnet (20240620) | 58.4 |
| 5 | GPT-4o Mini | 57.1 |
| 6 | Gemini 1.5 Pro (002) | 55.8 |
| 7 | Gemini 1.5 Flash (002) | 53.9 |
| 8 | InternVL3-78B | 52.8 |
| 9 | InternVL3-38B | 51.8 |
| 10 | Qwen 2 VL 72B | 51.3 |
| 11 | InternVL2.5-78B | 51.2 |
| 12 | Gemma 3 27B (IT) | 50.9 |
| 13 | Llama 4 Scout Base | 48.5 |
| 14 | Gemma 3 12B (IT) | 48.1 |
| 15 | InternVL3-8B | 45.4 |
Interactive version: theaggregate.ai/benchmark?slug=mega-bench-output-format-open-ended-output · How It Works · Data refreshed daily, snapshot 2026-09-05.