Video-Holmes: leaderboard
Complex video reasoning benchmark over suspense short films, requiring models to actively find and connect clues across video segments.
Metric: Overall (self-reported). Source: benchmarklist.com. Status: saturated. 16 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Pro | 45 |
| 2 | Gemini 2.0 Flash (Thinking) | 43.1 |
| 3 | GPT-4o | 42 |
| 4 | Gemini 1.5 Pro | 41.2 |
| 5 | Claude 3.7 Sonnet | 41 |
| 6 | Claude 3.5 Sonnet | 39.3 |
| 7 | InternVL3-8B | 32.3 |
| 8 | Gemini 2.0 Flash | 30.6 |
| 9 | O4 Mini | 29.9 |
Interactive version: theaggregate.ai/benchmark?slug=video-holmes · How It Works · Data refreshed daily, snapshot 2026-09-05.