OmniCap-IF - Instruction Satisfaction: leaderboard
Metric: Overall instruction satisfaction rate (ISR, %): share of instructions whose every constraint is satisfied, on multi-constraint captioning instructions on videos with sound; format constraints are checked by rule-based tools on LLM-extracted structure, content constraints by temporal-IoU tools or a gpt-5-mini judge answering binary and multiple-choice questions against the caption; higher is better. Source: arxiv.org. Saturation forecast: Around December 2027. 14 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 25.82 |
| 2 | Gemini 3 Flash | 23.55 |
| 3 | MiMo-V2.5 | 20.5 |
| 4 | MiMo-V2-Omni | 17.21 |
| 5 | MiniCPM-o-4.5 | 9.27 |
| 6 | Qwen3 Omni 30B A3B Instruct | 7.24 |
| 7 | Qwen2.5-Omni-7B | 2.34 |
Interactive version: theaggregate.ai/benchmark?slug=omnicap-if-instruction-satisfaction · How It Works · Data refreshed daily, snapshot 2026-09-29.