EgoCoT-Bench - Local Cause-Outcome Inference: leaderboard
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), Local Cause-Outcome Inference subtask (identify the recent action that caused an observed outcome); higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 19 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3.5 Plus | 74.72 |
| 2 | Qwen 3.5 122B A10B | 73.61 |
| 3 | Qwen 3.5 27B | 73.33 |
| 4 | GPT-5.2 | 72.22 |
| 5 | Qwen 3.5 397B A17B | 70.83 |
| 6 | GPT-5.1 | 70.56 |
| 7 | Qwen 3 VL 235B A22B | 62.5 |
Interactive version: theaggregate.ai/benchmark?slug=egocot-bench-local-cause-outcome-inference · How It Works · Data refreshed daily, snapshot 2026-10-07.