MEGA-Bench Task - Trance Physics Reasoning Basic — leaderboard

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

Top models

#ModelScore
1Gemini 2.5 Pro100
2GPT-4o88.2
3InternVL3-38B88.2
4InternVL3-14B88.2
5InternVL2.5-78B88.2
6Claude 3.5 Sonnet (20240620)88.2
7InternVL3-78B82.4
8InternVL3-8B82.4
9Gemini 1.5 Pro (002)76.5
10Qwen 2 VL 72B70.6
11Claude 3.5 Sonnet (20241022)70.6
12GPT-4o Mini70.6
13Gemini 2.0 Flash (Preview)70.6
14Gemma 3 27B (IT)64.7
15Gemma 3 12B (IT)47.1

Interactive version: theaggregate.ai/benchmark?slug=mega-bench-task-trance-physics-reasoning-basic · How the rankings work · Data refreshed daily, snapshot 2026-07-22.