AgenticVBench — leaderboard
Agentic video benchmark where autonomous agents perform multi-step video repurposing, sequencing, repair, and assembly tasks, scored by average task success.
Metric: Average Success (%). Source: agenticvbench.com. Status: saturation imminent. 10 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.6 Sol | 38.4 |
| 2 | Claude Fable 5 | 32.4 |
| 3 | GPT-5.5 | 31 |
| 4 | Gemini 3.1 Pro (Preview) | 23.8 |
| 5 | MiniMax-M3 | 22.7 |
| 6 | Claude Opus 4.7 | 22.1 |
| 7 | Gemini 3 Flash | 17.5 |
| 8 | Claude Sonnet 4.6 | 16.6 |
| 9 | GPT-5.4 Mini | 14.5 |
Interactive version: theaggregate.ai/benchmark?slug=agenticvbench · How the rankings work · Data refreshed daily, snapshot 2026-07-22.