EgoDyn-Bench: leaderboard
EgoDyn-Bench (2026) tests whether vision-language models understand the motion of the ego vehicle in driving video, through multiple-choice questions on speed, acceleration and turning answered from video alone. This board is balanced accuracy in percent on the semantic ego-motion questions; higher is better.
Metric: Balanced accuracy (%) of semantic ego-motion questions (14 templates of discrete choices on speed, acceleration and turning) from driving video alone, no explicit dynamic state inputs; higher is better. Source: arxiv.org. Saturation forecast: Around August 2027. 16 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 47 | #119 |
| 2 | GPT-5.1 | 45.2 | #240 |
| 3 | Cosmos-Reason2-8B | 39.9 | #755 |
| 4 | Qwen 3 VL 4B | 39.8 | #1151 |
| 5 | Qwen 3 VL 8B | 38.9 | #917 |
| 6 | InternVL3.5-8B | 38.8 | #1604 |
| 7 | RoboTron-Drive | 38.6 | |
| 8 | Claude Sonnet 4.5 | 38 | #261 |
| 9 | ImpromptuVLA (EgoDyn-Bench checkpoint unspecified) | 37.8 | |
| 10 | InternVL3.5-38B | 37.6 | #1021 |
| 11 | Qwen 3 VL 2B | 37.4 | #2134 |
| 12 | CamReasoner-8B | 36.9 | |
| 13 | Gemini 2.0 Flash | 36.5 | #723 |
| 14 | InternVL3.5-4B | 36 | #1930 |
| 15 | Cosmos-Reason2-2B | 35.1 | #1425 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=egodyn-bench · How It Works · Data refreshed daily, snapshot 2026-11-21.