EgoCoT-Bench - Reasoning Score: leaderboard
EgoCoT-Bench (2026) reasoning quality: on egocentric manipulation video questions, a Qwen-Max judge scores each model rationale from 0 to 5 against the annotated reference reasoning for coherence, factual consistency and grounding in the visual evidence. Mean score out of 5; higher is better.
Metric: Reasoning score (out of 5): a Qwen-Max judge rates each generated rationale against the annotated reference reasoning for logical coherence, factual consistency and alignment with the visual evidence, 0 to 5, mean over the 3,172 questions of EgoCoT-Bench; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 19 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3 VL Plus | 3.08 | #512 |
| 2 | Qwen 3.5 27B | 2.96 | #375 |
| 3 | Qwen 3 VL 32B | 2.96 | #622 |
| 4 | Qwen 3.5 122B A10B | 2.94 | #346 |
| 5 | Qwen 3.5 Plus | 2.92 | #242 |
| 6 | Qwen 3.5 397B A17B | 2.87 | #264 |
| 7 | GPT-5.2 | 2.85 | #185 |
| 8 | Qwen 3 VL 30B A3B | 2.79 | #784 |
| 9 | Qwen 3 VL 235B A22B | 2.78 | #485 |
| 10 | GPT-5.1 | 2.77 | #240 |
| 11 | Qwen 3 VL 8B | 2.73 | #917 |
| 12 | InternVL3.5-14B | 2.6 | #1504 |
| 13 | InternVL3.5-8B | 2.56 | #1604 |
| 14 | InternVL3.5-2B | 2.53 | #3284 |
| 15 | LLaVA-OneVision-1.5-4B | 2.5 | #1855 |
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-reasoning-score · How It Works · Data refreshed daily, snapshot 2026-11-21.