MEGA-Bench Task - Vibe Eval Short Phrase — leaderboard
Metric: Task Score (%). Source: huggingface.co. 27 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4o | 28.6 |
| 2 | Claude 3.5 Sonnet (20241022) | 28.6 |
| 3 | Gemini 1.5 Flash (002) | 28.6 |
| 4 | Claude 3.5 Sonnet (20240620) | 28.6 |
| 5 | Qwen 2 VL 72B | 21.4 |
| 6 | Pixtral-12B | 21.4 |
| 7 | InternVL2.5-78B | 21.4 |
| 8 | InternVL2-8B | 14.3 |
| 9 | Qwen 2 VL 2B | 14.3 |
| 10 | Gemini 1.5 Pro (002) | 14.3 |
| 11 | InternVL2.5-2B | 14.3 |
| 12 | Qwen 2 VL 7B | 7.1 |
| 13 | GPT-4o Mini | 7.1 |
| 14 | Aquila-VL-2B | 0 |
| 15 | MiniCPM-V-2.6 | 0 |
Interactive version: theaggregate.ai/benchmark?slug=mega-bench-task-vibe-eval-short-phrase · How the rankings work · Data refreshed daily, snapshot 2026-07-22.