MinCU: leaderboard
Metric: Grounded F1 (%; overall on 2,000 image pairs, 500 per change type; 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 | Qwen 3 VL 8B | 48.84 |
| 2 | Gemini 3.1 Pro (Preview) | 42.89 |
| 3 | Seed 1.8 | 42.43 |
| 4 | Gemini 2.5 Pro | 31.92 |
| 5 | GPT-5.4 | 19.58 |
Interactive version: theaggregate.ai/benchmark?slug=mincu · How It Works · Data refreshed daily, snapshot 2026-09-26.