4DCodeBench Shows Frontier Models Reconstruct Static Scenes but Fail at Dynamics
4DCodeBench · hf · 2026-10-06
4DCodeBench is a new benchmark testing agents on 4D inverse graphics through code generation: agents must reconstruct dynamic scenes from video as executable graphics programs, translating visual observations into compact representations of scene structure and dynamics, including physical simulations to reproduce complex behavior.
- Curated real-world videos plus synthetic scenes spanning deformation, fluid flow, and fracture.
- Extensive benchmarking of frontier models finds strong static reconstruction does not yet translate into reliable reconstruction of complex dynamics.
- The benchmark is open-sourced on GitHub as a testbed for agents that interpret world dynamics through code.
Related event: 4DCodeBench Tests Coding Agents on Rebuilding Dynamic 3D Scenes from Video(4 posts)→
More from Research
- Cambridge team's new work on pullback geometry for data on mixtures of manifolds — skoularidou · 2026-10-06
- CMU L3 Lab brings GradAlign RL data selection and sim2real papers to COLM 2026 — wellecks · 2026-10-06
- A statistical framework for LLM watermarks: optimal detection rules via hypothesis testing — weijie444 · 2026-10-06
- 4DCodeBench: new benchmark tests coding agents on reconstructing dynamic 3D scenes from video — CSProfKGD · 2026-10-06
- EMNLP paper: LLMs learn novel tasks more reliably from rules than from in-context examples — najoungkim · 2026-10-06
- New COLM study: Looped Transformers do implicit reasoning over parametric knowledge, boosting generalization — hhsun1 · 2026-10-06