MinCU - Relation: leaderboard
Metric: Grounded F1 (%; relation changes (an entity relocated with its identity preserved), 500 image pairs; a true positive needs the predicted box to reach IoU 0.5 with the changed region, the change description BERTScore 0.85 against the reference and the correct changed-entity label; zero-shot). Source: arxiv.org. Saturation forecast: Around 2033. 15 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 42.31 |
| 2 | Seed 1.8 | 40.85 |
| 3 | Qwen 3 VL 8B | 38.4 |
| 4 | GPT-5.4 | 18.36 |
Interactive version: theaggregate.ai/benchmark?slug=mincu-relation · How It Works · Data refreshed daily, snapshot 2026-09-26.