EgoCoT-Bench: leaderboard
The rapid development of Multimodal Large Language Models (MLLMs) has led to growing interest in egocentric video understanding, specifically the ability for MLLMs to recognize fine-grained hand-object interactions, track object state changes over time, and reason about manipulative processes in dynamic environments from a first-person perspective. However, existing egocentric video benchmarks suf
Metric: Mean (self-reported). Source: benchmarklist.com. Saturation forecast: Around July 2027. 20 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.5 27B | 71.28 | #187 |
| 2 | Qwen 3.5 Plus | 70.68 | #125 |
| 3 | Qwen 3.5 397B A17B | 70.11 | #146 |
| 4 | Qwen 3.5 122B A10B | 69.96 | #173 |
| 5 | GPT-5.2 | 67.91 | #110 |
| 6 | GPT-5.1 | 66.71 | #135 |
| 7 | Qwen 3 VL 235B A22B | 65.86 | #238 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=egocot-bench · How It Works · Data refreshed daily, snapshot 2026-10-20.