Progress-Bench - Vision-Based Demo PRC: leaderboard
Progress-Bench, from PROGRESSLM (Zhang et al., 2026), with the demonstration given as key frames labelled with their progress: the Spearman correlation (x100) between predicted and true progress along each robot manipulation trajectory, averaged over trajectories. Higher is better, from -100 to 100; direct predictions without prompted reasoning.
Metric: Progress rank correlation (Spearman x 100 between predicted and true progress along each trajectory, averaged over trajectories; answerable items with key-frame demonstrations; direct prediction). Source: arxiv.org. Saturation forecast: Estimated already saturated. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 89.4 | #163 |
| 2 | GPT-5 Mini | 87.7 | #349 |
| 3 | Qwen 2.5 VL 7B | 33.7 | #1796 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=progress-bench-vision-based-demo-prc · How It Works · Data refreshed daily, snapshot 2026-11-12.