OmniStarPro-Long - CDQ (Long Span): leaderboard
Cross-event difference queries: the model must report how a salient attribute such as a count, state or arrangement changed between two widely separated moments, on the OmniStarPro-Long partition of OmniStarPro (Yang et al., 2026), streams of 10 minutes to over an hour. Accuracy in percent judged by GPT-4o, on queries whose evidence lies more than 30 minutes before the query; higher is better.
Metric: Recall accuracy (%; share of cross-event difference (CDQ) queries, which ask how a salient attribute such as a count, state or spatial arrangement changed between two widely separated moments, that a GPT-4o judge marks as matching the ground truth; evidence over 30 minutes before the query; OmniStarPro-Long partition, streams of 10 minutes to over an hour). Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 5 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | LiveStarPro | 31.5 | |
| 2 | LiveStar | 17.7 | #1076 |
| 3 | MMDuet | 7.4 | |
| 4 | VideoLLM-MoD | 5.6 | |
| 5 | VideoLLM-online | 5.2 |
Interactive version: theaggregate.ai/benchmark?slug=omnistarpro-long-cdq-long-span · How It Works · Data refreshed daily, snapshot 2026-11-12.