EgoCoT-Bench - Next State Anticipation: leaderboard

EgoCoT-Bench (2026) Next State Anticipation subtask: from an egocentric manipulation video, predict the most likely next state of an object. Four-way multiple choice accuracy in percent; higher is better.

Metric: Answer accuracy (%, four-way multiple choice, strict exact match) of zero-shot MLLMs on egocentric manipulation videos of EgoCoT-Bench (351 clips, 3,172 human-verified questions built from spatio-temporal scene graphs), Next State Anticipation subtask (predict the most likely next state of an object); higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 19 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1GPT-5.186.99#240
2Qwen 3 VL 235B A22B85.97#485
3Qwen 3.5 Plus85.2#242
4Qwen 3.5 397B A17B84.95#264
5GPT-5.284.69#185
6Qwen 3 VL Plus84.69#512
7Qwen 3.5 27B84.65#375
8Qwen 3 VL 32B84.18#622
9Qwen 3 VL 8B83.03#917
10LLaVA-OneVision-1.5-8B82.14#1864
11Qwen 3 VL 30B A3B81.89#784
12Qwen 3.5 122B A10B81.63#346
13InternVL3.5-4B81.63#1930
14LLaVA-OneVision-1.5-4B81.38#1855
15InternVL3.5-8B80.36#1604

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=egocot-bench-next-state-anticipation · How It Works · Data refreshed daily, snapshot 2026-11-21.