PriVE-Bench: leaderboard
Metric: Counterfactual accuracy (%; macro-averaged over the five domains) on paired original and counterfactual images in five domains, counterfactual-only input, answers classified by a text-only judge (GPT-4o-mini for closed models, Qwen3-8B for open models); temperature 0, reasoning off or minimal. Source: arxiv.org. Saturation forecast: Around April 2027. 8 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5 (Minimal) | 46.7 |
| 2 | Claude Sonnet 4 | 46.6 |
| 3 | Qwen 3 VL 32B Instruct | 40.3 |
| 4 | Gemma 3 12B (IT) | 39.7 |
| 5 | Qwen 3 VL 8B Instruct | 31.6 |
| 6 | Gemma 3 27B (IT) | 30.5 |
| 7 | InternVL3.5-8B | 27.5 |
Interactive version: theaggregate.ai/benchmark?slug=prive-bench · How It Works · Data refreshed daily, snapshot 2026-09-29.