SpatialClaw: Training-Free Code-Action Agent Beats Prior by 11.2 Points on 20 Spatial Benchmarks
_akhaliq · x · 2026-09-30
NVIDIA researchers introduce SpatialClaw, a paper arguing that "code is the right action interface for spatial reasoning."
- Method: A VLM-backed agent writes Python in a persistent kernel, composing perception modules, inspecting intermediate results, and revising its strategy across steps. It is fully training-free, with no benchmark- or model-specific adaptation.
- Results: It beats a recent prior agent by +11.2 points across 20 benchmarks and improves consistently over six VLM backbones.
- Takeaway: Simply redesigning the agent's action interface around code execution yields large, general spatial-reasoning gains without any fine-tuning.
More from coding & agent
- Generic API-catalog MCP server vs one tool per job: a context-bloat experiment — MandeepKaur9407 · 2026-11-16
- Berkeley paper: LLMs answer with stale values even when the update is in context — ombre-ai · 2026-11-16
- skills-gate: one gate skill fronts 148 agent skills at just 389 tokens — Distinct-Pie2389 · 2026-11-16
- Open-weight Karpo-Decision-31B tops Decision Index at 73.9, one forward pass per answer — CodeByPoonam · 2026-11-16
- shadcn boosts @uisdk: one-shot UI rebuild that's instantly E2E-testable by AI — shadcn · 2026-11-16
- AI agent deletes 43,599 lines of OpenClaw test code in one PR — vincent_koc · 2026-11-16