MEGA-Bench Task - Visual Prediction Rater Depth Estimation: leaderboard

Metric: Task Score (%). Source: huggingface.co. 44 models tracked.

Top models

#ModelScore
1Gemini 2.5 Pro (Preview 03-25)78.6
2GPT-4o73.8
3Gemini 1.5 Pro (002)66.7
4Claude 3.5 Sonnet (20241022)59.5
5Claude 3.5 Sonnet (20240620)47.6
6InternVL3-78B45.2
7Gemma 3 27B (IT)42.9
8InternVL3-38B38.1
9Llama 4 Scout Base32.1
10InternVL2.5-78B31
11Qwen 2 VL 72B28.6
12Aquila-VL-2B28.6
13Gemma 3 4B (IT)23.8
14Gemini 1.5 Flash (002)21.4
15Qwen 2 VL 2B14.3

Interactive version: theaggregate.ai/benchmark?slug=mega-bench-task-visual-prediction-rater-depth-estimation · How It Works · Data refreshed daily, snapshot 2026-09-05.