LongVidSearch - State Mutation (3-Hop): leaderboard
Metric: Accuracy (%) on the 119 State Mutation questions that need exactly 3 evidence clips, judged against the reference by exact match or by majority vote of three LLM judges (GPT-5, Gemini 3 Pro, GPT-4o); the backbone drives a fixed VideoAgent-style framework that retrieves clips and reads their captions through a standard tool interface; higher is better. Source: arxiv.org. Saturation forecast: Around February 2028. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 36.13 | #91 |
| 2 | Gemini 3 Pro | 18.49 | #77 |
| 3 | Qwen 2.5 VL 72B Instruct | 17.65 | #364 |
| 4 | GPT-4o | 14.29 | #333 |
| 5 | Qwen 2.5 VL 7B Instruct | 7.69 | #643 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=longvidsearch-state-mutation-3-hop · How It Works · Data refreshed daily, snapshot 2026-10-11.