OmniCapBench - Ref Scene F1: leaderboard
Metric: Ref Scene F1 (%; structured audio-visual captioning, native atomic evaluation). Source: 01yzzyu.github.io. Saturation forecast: Around January 2027. 9 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Pro | 83.65 |
| 2 | Gemini 3.1 Pro (Preview) | 82.68 |
| 3 | Qwen 3.5 Omni Plus | 79.97 |
| 4 | Qwen3.5 Omni Flash | 66.07 |
Interactive version: theaggregate.ai/benchmark?slug=omnicapbench-ref-scene-f1 · How It Works · Data refreshed daily, snapshot 2026-10-10.