DRBench (Dense-Scene Reasoning) - Affordance Reasoning: leaderboard
Metric: Combined accuracy (%; object use within the scene context; 134 test questions). Source: arxiv.org. Saturation forecast: Around December 2026. 9 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Phi-4 Multimodal Instruct | 46.3 |
| 2 | InternVL2.5-2B | 35.8 |
Interactive version: theaggregate.ai/benchmark?slug=drbench-dense-scene-reasoning-affordance-reasoning · How It Works · Data refreshed daily, snapshot 2026-09-26.