PanoSpace-Bench - Relative 3D Position: leaderboard
Metric: Accuracy (%; 250 four-option multiple-choice questions on one ERP panorama). Source: arxiv.org. Saturation forecast: Around 2034. 12 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Ministral 3 14B | 31.6 |
| 2 | Ministral 3 8B | 29.6 |
| 3 | Qwen 2.5 VL 7B Instruct | 29.6 |
| 4 | GPT-4o | 27.2 |
| 5 | Qwen 3.5 9B | 24.5 |
| 6 | MiMo-V2.5 | 23.6 |
| 7 | Qwen 3 VL 8B | 23.6 |
Interactive version: theaggregate.ai/benchmark?slug=panospace-bench-relative-3d-position · How It Works · Data refreshed daily, snapshot 2026-09-25.