MultiHaystack - Video: leaderboard
Metric: Answer accuracy (%) on the 105 questions whose evidence is a video, where the model answers from the top-5 items that the E5-V retriever returns from MultiHaystack's 46,260-item mixed corpus of images, videos and documents, graded correct or incorrect by a GPT-4o-mini judge; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 7 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 60 | #91 |
| 2 | Gemini 2.5 Flash | 52.38 | #237 |
| 3 | InternVL3-8B | 17.14 | #606 |
| 4 | Qwen 2 VL 7B | 16.19 | #741 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=multihaystack-video · How It Works · Data refreshed daily, snapshot 2026-10-11.