EgoSafetyBench - Visual-Channel Mismatch Detection: leaderboard

Metric: F1 (%) of flagging chunks whose in-scene sign, sticker or label misrepresents the physical situation, pooled over the chunks of the 400-scenario visual-channel track of paired misleading and truthful signs; ten VLM safety guards judging egocentric robot videos split into ten half-second chunks (ten frames each at 1280x720), each chunk judged alone under the current-chunk protocol with a single-word verdict and greedy decoding; higher is better. Source: arxiv.org. Saturation forecast: Around December 2026. 10 models tracked.

Top models

#ModelScore
1Claude Sonnet 4.685
2Gemini 3.5 Flash82.1
3Gemini 3.1 Flash Lite80.5
4Qwen 3.5 4B67.8
5Qwen 3 VL 4B Instruct66.3
6Gemma 3 4B (IT)53
7Qwen 3.5 0.8B39
8InternVL3.5-8B36.8

Interactive version: theaggregate.ai/benchmark?slug=egosafetybench-visual-channel-mismatch-detection · How It Works · Data refreshed daily, snapshot 2026-09-29.