SpatialScout benchmark tests VLM agents on active 3D mapping across 80 scenes
Zhaoxi Zhang · hf · 2026-10-22
Researchers from Peking University and Tianjin University introduce SpatialScout, a benchmark evaluating vision-language model agents on active object-level 3D map building. It spans 80 indoor scenes with 3,140 target objects: given an object inventory, agents must autonomously explore, gather RGB observations with camera calibration and poses, and integrate them into a coherent metric map. Results reveal major gaps in precise metric map construction, showing spatial intelligence remains an open challenge beyond passive question answering.
More from Research
- 130K-sample RAG claim verification dataset lands on Hugging Face — MaziyarPanahi · 2026-10-22
- Tencent Hunyuan's WorkForge builds 16.7K verifiable real-world environments for work agents — TencentHunyuan · 2026-10-22
- Gauss vs the machine: what the 1-to-100 sum trick says about ML and computation — tobias_rees · 2026-10-22
- Radiolab's 'Math Vs Machine': OpenAI claims a Millennium Prize Problem solve — stevenstrogatz · 2026-10-22
- UK/Nigeria driver-coaching NLG study wins Best Evaluation award at INLG 2026 — EhudReiter · 2026-10-22
- 31,000-question study finds LLMs across 24 languages are secretly obsessed with Japan — dioscuri · 2026-10-22