CausalPhys - Anticipation: leaderboard

Metric: Answer accuracy (0-1 scaled to %) on the Anticipation domain (predicting physical outcomes such as collisions, deformation and fluid flow) of the image- and video-based questions with verified answers over real-world physical scenes; higher is better. Source: arxiv.org. Saturation forecast: Around 2031. 11 models tracked.

Top models

#ModelScore
1GPT-4o60.11
2GPT-4o Mini59.11
3Gemini 2.5 Flash58.22
4InternVL3-78B58
5Phi-4 Multimodal Instruct55.33
6Claude Sonnet 453.22
7Qwen 2 VL 7B52.22
8Qwen 3 VL 32B51.89
9Mistral Small 3.241

Interactive version: theaggregate.ai/benchmark?slug=causalphys-anticipation · How It Works · Data refreshed daily, snapshot 2026-09-29.