Embodied3DBench - Affordance (3D): leaderboard
Metric: Affordance (3D) task (functional contact point and interaction vector in metric 3D), normalized score (0-100): IoU of predicted boxes or the L2 error of predicted points (pixels in 2D, metres in 3D) mapped to 0-100, on the simulated robotic tabletop scenes of Embodied3DBench (21K QA pairs over 4.8K object instances), zero-shot; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 13 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Pro | 54.8 |
| 2 | GPT-5 | 53.1 |
| 3 | Claude Sonnet 4 | 48.4 |
| 4 | Gemini 2.5 Flash | 43 |
| 5 | Doubao-Seed-1.6 (Thinking) | 39.5 |
| 6 | GPT-4o | 30.7 |
| 7 | Qwen 3 VL 4B Instruct | 28.5 |
| 8 | InternVL3-38B | 23.9 |
| 9 | Qwen 2.5 VL 72B Instruct | 22.2 |
| 10 | Qwen 2.5 VL 7B Instruct | 20 |
Interactive version: theaggregate.ai/benchmark?slug=embodied3dbench-affordance-3d · How It Works · Data refreshed daily, snapshot 2026-10-07.