CapRiCorn-1K - Referential Consistency: leaderboard
Metric: Subject referential consistency (%; share of same-subject description pairs a GPT-4.1 judge clusters as one referent, over all pairs of that subject's keypoints, audiovisual captions of 1,000 videos). Source: arxiv.org. Saturation forecast: Around December 2026. 18 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Flash | 39.6 |
| 2 | Gemini 3.1 Pro (Preview) | 39.1 |
| 3 | Qwen3 Omni 30B A3B Instruct | 1.6 |
| 4 | Qwen2.5-Omni-7B | 0.6 |
Interactive version: theaggregate.ai/benchmark?slug=capricorn-1k-referential-consistency · How It Works · Data refreshed daily, snapshot 2026-09-26.