MEGA-Bench Task - Code Programming Test Easy: leaderboard
Metric: Task Score (%). Source: huggingface.co. 44 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | InternVL2.5-78B | 56.9 |
| 2 | InternVL3-38B | 55.6 |
| 3 | Gemini 2.5 Pro (Preview 03-25) | 54.2 |
| 4 | Claude 3.5 Sonnet (20240620) | 54.2 |
| 5 | InternVL3-78B | 50.7 |
| 6 | GPT-4o | 45.8 |
| 7 | Claude 3.5 Sonnet (20241022) | 45.8 |
| 8 | Gemma 3 27B (IT) | 43.8 |
| 9 | Qwen 2 VL 72B | 27.1 |
| 10 | Llama 4 Scout Base | 27.1 |
| 11 | InternVL3-8B | 25.7 |
| 12 | Qwen 2 VL 7B | 20.8 |
| 13 | GPT-4o Mini | 20.8 |
| 14 | Pixtral-12B | 20.8 |
| 15 | Gemini 1.5 Flash (002) | 18.8 |
Interactive version: theaggregate.ai/benchmark?slug=mega-bench-task-code-programming-test-easy · How It Works · Data refreshed daily, snapshot 2026-09-05.