EmbodiedEvo: Self-improving Coding Agents for Physical Intelligence

A team led by Wenyi Li submitted EmbodiedEvo, a self-improving coding agent for physical tasks, to arXiv on 13 Oct 2026 (latest version 14 Oct 2026), reporting an average score of 44.48 and a 37.82% success rate on RoboDojo-Sim, surpassing all officially verified results. The system evolves reusable Skills and Lessons through feedback on a single Playground scene, then freezes the resulting harness for new scenes; experience from the stronger Astra model raised Sol and Luna's success rates by 8.86 and 5.50 percentage points over no experience, and Skills and Lessons achieved comparable task performance at 53.0% lower execution cost than raw interaction history. Paired tests showed the largest gains on memory-dependent tasks and successful reuse of coordinated procedures, while fine contact control remained difficult.

Computer Science Robotics Submitted on 13 Oct 2026 this version , latest version 14 Oct 2026 v2 https://arxiv.org/abs/2610.18867v2 Title:EmbodiedEvo: Self-improving Coding Agents for Physical Intelligence View PDF http://arxiv.org/pdf/2610.18867v1 HTML experimental https://arxiv.org/html/2610.18867v1 Abstract:Bringing general-purpose agents into the physical world calls for an executable representation of physical intelligence. Compared with vision-language-action VLA models, which map observations and language instructions directly to actions, code makes control logic explicit and reusable. Code can serve not only as a policy representation but also as the workspace in which the agent solves a task. Rather than following a prescribed workflow of planning, execution, and verification, the agent can autonomously inspect the state, revise and execute code, and use feedback to choose the next step. The success of this SWE-style loop in digital environments motivates its extension to physical tasks. Yet physical execution remains prone to recurring failures. To retain and refine experience for later scenes, we introduce EmbodiedEvo, where a coding agent evolves reusable Skills and Lessons through feedback on a single Playground scene, then freezes the resulting harness for diverse new scenes. On RoboDojo-Sim, EmbodiedEvo achieves an average score of 44.48 and a success rate of 37.82\%, surpassing all officially verified results. Paired tests show the largest gains on memory-dependent tasks and successful reuse of coordinated procedures, while fine contact control remains difficult. Experience from stronger models also improves weaker executors: Astra's experience raises Sol and Luna's success rates by 8.86 and 5.50 percentage points over no experience. Compared with raw interaction history, Skills and Lessons achieve comparable task performance at 53.0\% lower execution cost. Submission history From: Wenyi Li view email http://arxiv.org/show-email/0abc43c8/2610.18867 v1 Tue, 13 Oct 2026 17:59:51 UTC 13,771 KB Wed, 14 Oct 2026 17:45:55 UTC 13,742 KB \ v2\ http://arxiv.org/abs/2610.18867v2 Additional Features References & Citations Loading... Bibliographic and Citation Tools Bibliographic Explorer What is the Explorer? https://info.arxiv.org/labs/showcase.html arxiv-bibliographic-explorer Connected Papers What is Connected Papers? https://www.connectedpapers.com/about Litmaps What is Litmaps? https://www.litmaps.co/ scite Smart Citations What are Smart Citations? https://www.scite.ai/ Code, Data and Media Associated with this Article alphaXiv What is alphaXiv? https://alphaxiv.org/ CatalyzeX Code Finder for Papers What is CatalyzeX? https://www.catalyzex.com DagsHub What is DagsHub? https://dagshub.com/ Gotit.pub What is GotitPub? http://gotit.pub/faq Hugging Face What is Huggingface? https://huggingface.co/huggingface ScienceCast What is ScienceCast? https://sciencecast.org/welcome Demos Recommenders and Search Tools Influence Flower What are Influence Flowers? https://influencemap.cmlab.dev/ CORE Recommender What is CORE? https://core.ac.uk/services/recommender arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs https://info.arxiv.org/labs/index.html .