LongVideoBench: leaderboard
Long-form video understanding benchmark requiring models to reason over extended temporal contexts.
Metric: Test Total Accuracy (%). Source: longvideobench.github.io. Saturation forecast: Estimated already saturated. 35 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 1.5 Pro | 64.4 | #445 |
| 2 | Gemini 1.5 Flash | 62.4 | #618 |
| 3 | GPT-4 Turbo | 60.7 | #404 |
| 4 | GPT-4o Mini | 58.8 | #599 |
| 5 | MiniCPM-V-2.6 | 57.7 | #836 |
| 6 | Qwen 2 VL 7B | 56.8 | #750 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=longvideobench · How It Works · Data refreshed daily, snapshot 2026-10-20.