Code World Model uses coding agents as the world brain
mark_k · x · 2026-08-28
Introduces a novel "Code World Model" architecture that separates world evolution from visual rendering by using a coding agent as the world brain.
Key aspects of the framework:
- Coding Agent: Acts as the reasoning engine, writing executable code to maintain persistent world state, rules, and consequences.
- Video Model: Controlled by a visual proxy compiled from the code state, responsible for rendering high-fidelity visuals.
This approach addresses the limitations of video-based models in maintaining long-term consistency and memory. Demonstrations using fine-tuned MiniMax-H3 show interactive worlds generated under spatial constraints derived from the coding agent.
Related event: Code World Model: Westlake University Proposes Coding Agent as World Brain(5 posts)→
More from Research
- WorldMemArena Accepted at NeurIPS: Memory Storage ≠ Memory Use — xwang_lk · 2026-10-16
- Mech Interp Veteran Rethinks Alignment, Misgeneralization, CoT Oversight — akbirthko · 2026-10-16
- WEDGE paper generates training environments conditioned on navigation agent failures — _akhaliq · 2026-10-16
- RExBench: best coding agent completes only 33% of autonomous research extensions — najoungkim · 2026-10-16
- Google's CLASHarness benchmarks multi-agent conflicts in computer use, finds frontier VLMs struggle — google · 2026-10-16
- Tsinghua team proves vanilla flow-matching sampling can't exceed constant-depth expressive power — SQZ · 2026-10-16