ChronoPhyBench - Next-State Selection (6 Options): leaderboard

Metric: Accuracy (%) on ChronoPhyBench next-state frame selection with 6 candidates: given a real physical-interaction video context (8 uniformly sampled frames) and its caption, choose the frame that comes next; closed models at temperature 0.1, open models at their recommended sampling; chance 16.7; higher is better. Source: arxiv.org. Saturation forecast: Around December 2026. 25 models tracked.

Top models

#ModelScore
1GPT-5.570.5
2Doubao-Seed-2.0-Pro-26021565
3Gemini 3 Pro (Preview)63
4Doubao-Seed-1-8-25122853
5Gemini 2.5 Pro46.63
6GLM-4.1V-9B (Thinking)39.5
7Doubao-Seed-1.6-thinking-25071538.5
8Gemini 2.5 Flash37.85
9GPT-5.437.5
10GPT-5.334.5
11Qwen 3 VL 32B Instruct25.5
12Qwen 3 VL 8B (Thinking)25
13InternVL3-8B23.5
14Qwen 3 VL 8B Instruct22.5
15InternVL3.5-8B19.5

Interactive version: theaggregate.ai/benchmark?slug=chronophybench-next-state-selection-6-options · How It Works · Data refreshed daily, snapshot 2026-09-29.