OVO-S-Bench - Egocentric Perception: leaderboard

Metric: Level 1 (instantaneous egocentric perception) multiple-choice accuracy (%) on the 1,680-question evaluation set, prefix-access mode (128 frames sampled uniformly from the video before the query timestamp, question given with the frames), mean of its three task families. Source: arxiv.org. Saturation forecast: Around 2031. 31 models tracked.

Top models

#ModelScore
1Gemini 3.1 Pro (Preview)61.92
2GPT-5.454.64
3Gemini 3.1 Flash Lite54.09
4Qwen 3 VL 235B A22B52.45
5Qwen 3.5 27B (Non-reasoning)51.5
6Qwen 3 VL 32B Instruct50.14
7Qwen 3.5 397B A17B49.57
8Gemma 4 26B A4B49.33
9Qwen 3.5 9B (Non-reasoning)47.47
10InternVL3.5-8B45.92
11Qwen 3.5 4B (Non-reasoning)45.41
12Grok 4.1 Fast (Non-reasoning)44.8
13Qwen 3 VL 4B Instruct43.25
14Gemma 4 E4B40.92
15Qwen 2.5 VL 7B Instruct40.67

Interactive version: theaggregate.ai/benchmark?slug=ovo-s-bench-egocentric-perception · How It Works · Data refreshed daily, snapshot 2026-09-29.