{"slug": "training-reasoning-models-in-simulation-to-natively-use-real-robots", "title": "Training Reasoning Models in Simulation to Natively Use Real Robots", "summary": "An open-weight 27B reasoning model trained entirely in simulation with reinforcement learning succeeded in 60% of trials on a real dual-arm robot, where the base model succeeded in none, according to an arXiv paper submitted on 13 November 2026. Using a Post-training Engine for Agentic Robotics (PEAR) and robot-use reinforcement learning (RURL) on LIBERO and RoboDojo, the training raised LIBERO success from 1.6% to 74.4%, approaching GPT-6 Astra and GPT-6.1 Sol at maximum reasoning effort, and reached 57.4% on LIBERO-Pro versus 0.75% for the demonstration-fine-tuned vision-language-action model π0.5. The model retained its general capabilities with no measurable degradation on nine reasoning, coding and vision benchmarks, indicating simulation-trained robot-use reinforcement learning can transfer to unseen robots, the real world and novel instructions.", "body_md": "# Computer Science > Robotics\n\n  [Submitted on 13 Nov 2026]\n\n# Title:Training Reasoning Models in Simulation to Natively Use Real Robots\n\n[View PDF](http://arxiv.org/pdf/2611.18562v1)\n\n[HTML (experimental)](https://arxiv.org/html/2611.18562v1)\n\nAbstract:Robot-use agents (RUAs) have emerged as a promising approach to general-purpose robot control. However, strong performance has so far been achieved only by frontier proprietary models such as GPT-6 Astra and Claude Fable 5.2, whereas open-weight models remain substantially less capable. We show that an open-weight 27B reasoning model can be trained to use real robots natively through reinforcement learning, entirely in simulation and without demonstrations. Like a computer-use agent (CUA), the model observes the environment, reasons in a chain of thought, and generates the next end-effector pose, without any harness. We build a Post-training Engine for Agentic Robotics (PEAR) and use it to train the model on simulated robot tasks at scale with robot-use reinforcement learning (RURL), mirroring how frontier models acquire computer use. We find that the skills acquired in simulation generalize well beyond the training tasks: they transfer to an unseen robot, to the real world, and to novel instructions. After training on LIBERO and RoboDojo in simulation, the model succeeds in 60% of our trials on a real dual-arm robot, whereas the base model succeeds in none. On LIBERO, training raises the success rate from 1.6% to 74.4%, approaching GPT-6 Astra and GPT-6.1 Sol at maximum reasoning effort. Compared with $\\pi_{0.5}$, a vision-language-action model fine-tuned on demonstrations, our model is substantially more robust to changes in the task itself. On LIBERO-Pro, $\\pi_{0.5}$ succeeds in only 0.75% of episodes when assigned a new task in a familiar scene whereas our model succeeds in 57.4%. Finally, the model retains its general capabilities, with no measurable degradation on nine reasoning, coding and vision benchmarks. These results highlight robot-use reinforcement learning as a promising component of foundation-model post-training.\n    \n\n### Additional Features\n\n### References & Citations\n\nLoading...\n\n# Bibliographic and Citation Tools\n\nBibliographic Explorer \n\n*(*[What is the Explorer?](https://info.arxiv.org/labs/showcase.html#arxiv-bibliographic-explorer))\nConnected Papers \n\n*(*[What is Connected Papers?](https://www.connectedpapers.com/about))\nLitmaps \n\n*(*[What is Litmaps?](https://www.litmaps.co/))\nscite Smart Citations \n\n*(*[What are Smart Citations?](https://www.scite.ai/))\n# Code, Data and Media Associated with this Article\n\nalphaXiv \n\n*(*[What is alphaXiv?](https://alphaxiv.org/))\nCatalyzeX Code Finder for Papers \n\n*(*[What is CatalyzeX?](https://www.catalyzex.com))\nDagsHub \n\n*(*[What is DagsHub?](https://dagshub.com/))\nGotit.pub \n\n*(*[What is GotitPub?](http://gotit.pub/faq))\nHugging Face \n\n*(*[What is Huggingface?](https://huggingface.co/huggingface))\nScienceCast \n\n*(*[What is ScienceCast?](https://sciencecast.org/welcome))\n# Demos\n\n# Recommenders and Search Tools\n\nInfluence Flower \n\n*(*[What are Influence Flowers?](https://influencemap.cmlab.dev/))\nCORE Recommender \n\n*(*[What is CORE?](https://core.ac.uk/services/recommender))\n# arXivLabs: experimental projects with community collaborators\n\narXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.\n\nBoth 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.\n\nHave an idea for a project that will add value for arXiv's community? [**Learn more about arXivLabs**](https://info.arxiv.org/labs/index.html).", "url": "https://wpnews.pro/news/training-reasoning-models-in-simulation-to-natively-use-real-robots", "canonical_source": "http://arxiv.org/abs/2611.18562v1", "published_at": "2026-11-16 15:43:45+00:00", "updated_at": "2026-11-16 16:23:18.177892+00:00", "lang": "en", "topics": ["robotics", "ai-research", "machine-learning", "large-language-models", "ai-agents"], "entities": ["PEAR", "RURL", "LIBERO", "RoboDojo", "LIBERO-Pro", "GPT-6 Astra", "GPT-6.1 Sol", "π0.5"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/training-reasoning-models-in-simulation-to-natively-use-real-robots", "markdown": "https://wpnews.pro/news/training-reasoning-models-in-simulation-to-natively-use-real-robots.md", "text": "https://wpnews.pro/news/training-reasoning-models-in-simulation-to-natively-use-real-robots.txt", "jsonld": "https://wpnews.pro/news/training-reasoning-models-in-simulation-to-natively-use-real-robots.jsonld"}}