MEGA-Bench Task - CLEVR Arithmetic: leaderboard
Metric: Task Score (%). Source: huggingface.co. 44 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Pro (Preview 03-25) | 78.9 |
| 2 | Qwen 2 VL 72B | 68.4 |
| 3 | InternVL2.5-78B | 68.4 |
| 4 | Llama 4 Scout Base | 68.4 |
| 5 | GPT-4o | 63.2 |
| 6 | Claude 3.5 Sonnet (20240620) | 63.2 |
| 7 | InternVL3-78B | 52.6 |
| 8 | Qwen 2 VL 7B | 52.6 |
| 9 | Claude 3.5 Sonnet (20241022) | 52.6 |
| 10 | InternVL3-38B | 52.6 |
| 11 | InternVL2-8B | 47.4 |
| 12 | InternVL3-8B | 42.1 |
| 13 | Qwen 2 VL 2B | 36.8 |
| 14 | Gemini 1.5 Pro (002) | 36.8 |
| 15 | Pixtral-12B | 36.8 |
Interactive version: theaggregate.ai/benchmark?slug=mega-bench-task-clevr-arithmetic · How It Works · Data refreshed daily, snapshot 2026-09-05.