MEGA-Bench Task - SciBench With Solution Open Ended — leaderboard

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

Top models

#ModelScore
1Gemini 2.5 Pro72.6
2Gemini 2.0 Flash (Preview)61.8
3Gemini 1.5 Pro (002)48
4Gemini 1.5 Flash (002)45
5Gemma 3 27B (IT)40.8
6InternVL3-38B38.8
7InternVL2.5-78B38.6
8InternVL3-78B36.4
9GPT-4o36.2
10GPT-4o Mini35.6
11Gemma 3 12B (IT)33.8
12Qwen 2 VL 72B33.8
13Llama 4 Scout Base33.4
14Ovis2-8B32.4
15InternVL3-14B32

Interactive version: theaggregate.ai/benchmark?slug=mega-bench-task-scibench-with-solution-open-ended · How the rankings work · Data refreshed daily, snapshot 2026-07-22.