VSAS-Bench (Synchronous) - Present: leaderboard

VSAS-Bench (2026) Present tasks, questions about events happening now in a streamed video, under the synchronous protocol; per-timestep responses judged against dense reference annotations, accuracy (%), higher is better.

Metric: Mean average accuracy (%) on Present tasks (queries about events happening now) under the synchronous protocol (the model runs in lockstep with the stream, so inference latency does not count; video VLMs adapted with the sliding-window memory policy), over VSAS-Bench's 92 videos with dense 1 FPS free-form annotations; each timestep's response is judged against the annotation by GPT-5 (medium reasoning) on a 0-1 scale, shown as %; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 12 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1GPT-5Low83.1#178
2Qwen 3 VL 32B Instruct69.4#606
3Qwen 3 VL 4B Instruct64.3#1332
4InternVL3-38B64.1#939
5Qwen 3 VL 8B Instruct61.7#1029
6InternVL3-8B59.2#1782
7Qwen 3 VL 2B Instruct54.3#2367
8InternVL3-2B52.8#3290
9StreamBridge [VSAS-Bench checkpoint unspecified]50.1
10Dispider48.5#3815
11VideoLLM-Online20
12Flash-VStream-7B19.8

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=vsas-bench-synchronous-present · How It Works · Data refreshed daily, snapshot 2026-11-21.