UC Berkeley Researchers Release CUA-Lite, an Open Platform Unifying Sandboxes, Data, Evaluation and RL for Computer-Use Agents

UC Berkeley researchers released CUA-Lite, an open platform unifying sandboxes, data, evaluation, and reinforcement learning for computer-use agents, with Lite.OSWorld reproducing the OSWorld task suite in Docker containers instead of KVM virtual machines, cutting memory from 4.1 GB to 0.9 GB and enabling ~4.6× more parallel instances while matching OSWorld scores across 13 models. The platform includes a unified LiteSample schema with 10+ preprocessed datasets on Hugging Face, 10+ integrated agents, and 15+ benchmarks, totaling 30k+ verifiable tasks.

A team of researchers from UC Berkeley have released CUA-Lite https://cua-lite.github.io/ , an open platform for computer-use agents CUAs . The argument behind it is infrastructural rather than model-centric: training and benchmarking a CUA requires four pieces: agents, environments, traces, and a framework to evaluate and train them and all four are currently fragmented across separate repositories with incompatible interfaces. CUA-Lite puts them behind one action space, one data schema, and one command, across desktop, browser and mobile. Is it deployable? Yes. The stack installs with uv sync --all-extras on Python 3.12, and its lightweight sandboxes run on any Docker host without /dev/kvm , so cloud instances, CI runners and nested containers all work. The VM tax, and how Lite.OSWorld removes it The most concrete contribution is Lite.OSWorld https://cua-lite.github.io/blog/kvm-free-osworld/ . OSWorld https://github.com/xlang-ai/OSWorld provides a faithful Ubuntu desktop, but it ships as a full QEMU/KVM virtual machine per task, requiring nested virtualization that most managed infrastructure does not expose. CUA-Lite reproduces the same task suite and the same evaluators on a GNOME desktop inside a plain Docker container. | Task | OSWorld | Lite.OSWorld | |---|---|---| | Runtime | QEMU/KVM VM | Docker container | | Host requirement | /dev/kvm , nested virt | Any Docker host | | Memory | 4.1 GB | 0.9 GB | | Cold start | 29.9 s | 23.8 s | | Parallelism | baseline | ~4.6× more instances | | Task suite | OSWorld | Identical | Fidelity is the obvious concern when you swap a VM for a container, and the team addresses it directly: across 13 models, Lite.OSWorld scores match the OSWorld VM’s, so a score or a training signal earned in the container transfers back to the real benchmark. The same base now carries a family of sandboxes: Lite.ScaleCUA, Lite.CUAGym and Lite.CUAWorld, the last expanding into roughly 40 applications including Blender, QGIS and VS Code. In total the platform claims 30k+ verifiable tasks. One schema for data, one adapter per model CUA-Lite’s second layer is LiteSample , a single supervised-learning schema shared across every environment, agent and task type, shipped as plain parquet plus images. Ten-plus existing CUA datasets have been preprocessed into it and published free on Hugging Face https://huggingface.co/cua-lite , including Aguvis https://huggingface.co/datasets/cua-lite/Aguvis , OpenCUA https://huggingface.co/datasets/cua-lite/OpenCUA , ScaleCUA https://huggingface.co/datasets/cua-lite/ScaleCUA , GUI-360 https://huggingface.co/datasets/cua-lite/GUI-360 , GUIOdyssey https://huggingface.co/datasets/cua-lite/GUIOdyssey and Multimodal-Mind2Web https://huggingface.co/datasets/cua-lite/Multimodal-Mind2Web . Alongside those corpora sit fresh rollout datasets https://huggingface.co/collections/cua-lite/rollouts generated by rolling a frontier teacher model through the sandboxes, for distillation into smaller students. Because model families expect different scaffolding, the framework ships a per-model adapter that packs a unified LiteSample into each model’s own training format, including history collapsing so several steps share one forward pass. Eval, SFT and RL behind one command Agents and environments meet in lite.gym : screenshots up, actions down, with one action space per platform. 10+ agents are built in GPT, Claude, Gemini, Qwen3-VL, UI-TARS https://huggingface.co/ByteDance-Seed/UI-TARS-1.5-7B , Fara-7B https://huggingface.co/microsoft/Fara-7B , MAI-UI and others, and 15+ benchmarks are integrated, spanning grounding ScreenSpot-Pro https://github.com/cua-lite/cua-lite/blob/main/lite/gym/envs/screenspot pro/README.md , OSWorld-G , desktop OSWorld, OSWorld-2, WindowsAgentArena, CUABench , browser WebArena https://github.com/web-arena-x/webarena , VisualWebArena, MiniWoB, WebVoyager, Online-Mind2Web, WebGym and mobile AndroidWorld https://github.com/google-research/android world , AndroidLab, MobileWorld, MobileGym . Swapping --model-id and --env-id in scripts/rollout.py is the whole interface. The same loop serves training. For SFT, the README https://github.com/cua-lite/cua-lite documents fine-tuning Qwen3-VL-2B-Instruct on Lite.ScaleCUA https://huggingface.co/datasets/cua-lite/Lite.ScaleCUA desktop trajectories, lifting mean episode return from 0.138 to 0.237 on the 332-task lite.osworld eval split, a single reported configuration on two GPUs, not an independently reproduced result. For RL, rollouts scored in the environment drive GRPO updates on top of Slime, with a worked MobileGym example covering 416 mobile tasks across 28 apps. Interactive explainer Key Takeaways - CUA-Lite unifies agents, environments, traces and training under one action space and one LiteSample schema. - Lite.OSWorld runs OSWorld tasks VM-free in Docker at 0.9 GB versus 4.1 GB, roughly 4.6× more parallel desktops. - Scores in the container match the OSWorld VM across 13 models, so training signal transfers to the real benchmark. - 30k+ verifiable tasks, 15+ benchmarks, 10+ agents, and 20+ datasets published free on Hugging Face. - Deployable on any Docker host, but the repository ships no explicit license yet — verify terms before commercial use. Check out the Project Page https://cua-lite.github.io/ , GitHub Repo https://github.com/cua-lite/cua-lite and Datasets on Hugging Face https://huggingface.co/cua-lite . Also, feel free to follow us on Twitter https://x.com/intent/follow?screen name=marktechpost and don’t forget to join our 150k+ML SubReddit https://www.reddit.com/r/machinelearningnews/ and Subscribe to our Newsletter https://magic.beehiiv.com/v1/f5e63dd4-5653-4f09-83e2-321a8b1ba526?email={{email}} . Wait are you on telegram? now you can join us on telegram as well. https://t.me/machinelearningresearchnews Need to partner with us for promoting your GitHub Repo OR Hugging Face Page OR Product Release OR Webinar etc.? Connect with us https://forms.gle/wbash1wF6efRj8G58 Asif Razzaq is the CEO of Marktechpost Media Inc.. As a visionary entrepreneur and engineer, Asif is committed to harnessing the potential of Artificial Intelligence for social good. His most recent endeavor is the launch of an Artificial Intelligence Media Platform, Marktechpost, which stands out for its in-depth coverage of machine learning and deep learning news that is both technically sound and easily understandable by a wide audience. The platform boasts of over 2 million monthly views, illustrating its popularity among audiences.