OpenCUA

Visit Tool

OpenCUA is an AI Agents & Automation tool that provides open foundations for computer-use agents. It includes large-scale datasets, models, and evaluation benchmarks for developing computer-use AI.

Claim this tool

1View

At a glance

Pricing

Open Source

Free tier

Yes

API

Skill level

Technical

About

What is OpenCUA?

OpenCUA is a comprehensive open-source framework designed for scaling Computer-Use Agent (CUA) data and foundation models. It features AgentNet, the first large-scale computer-use task dataset spanning multiple operating systems and applications, and AgentNetTool, an annotation infrastructure for capturing human computer-use demonstrations. The framework also includes AgentNetBench, an offline evaluator for benchmarking model-predicted actions, and OpenCUA Models, end-to-end computer-use foundation models with strong planning and grounding capabilities. Notably, OpenCUA-72B achieves state-of-the-art performance on OSWorld-Verified, making it ideal for developers and researchers working on advanced AI agents.

Best used for

Ideal for developers and data scientists who need to build, train, and evaluate computer-use AI agents across different operating systems. Especially valuable for researchers and teams looking to leverage large-scale datasets and state-of-the-art foundation models for agent development and benchmarking.

Common actions

develop AI agents

benchmark agent performance

process computer-use data

train foundation models

automated workflowworkflowsopen-sourcedeepfakecollaborationlow-code/no-codegithub copilot"AI Agents"face swapping

Capabilities

Key features

Large-scale computer-use dataset
Human demonstration annotation
Offline agent evaluator
End-to-end foundation models
Multimodal Rotary Position Embedding
vLLM support

Target Audience

developerdata scientiststartup founder

Integrations

Not yet documented

Pricing & Plans

Open Source

Free

FAQs

What is AgentNet and how can I access it?

AgentNet is the first large-scale desktop computer-use agent trajectory dataset, containing 22.6K human-annotated tasks across Windows, macOS, and Ubuntu. You can download the dataset from Hugging Face using the provided pip install and huggingface-cli commands, then unzip the files for use.

How does OpenCUA support model deployment and inference?

OpenCUA models are fully supported by vLLM for production deployment, including OpenCUA-7B, OpenCUA-32B, and OpenCUA-72B. It also supports inference with HuggingFace Transformers. Specific commands and examples are provided for serving models with vLLM and running grounding examples.

What is the purpose of AgentNetBench?

AgentNetBench is an offline evaluator that provides a realistic way to benchmark OS agent trajectories. It compares model-predicted low-level actions against ground-truth human actions, reporting detailed metrics to assess agent performance and identify areas for improvement.

Trending

Subcategories trending in AI Agents & Automation

Chatbots & Conversational AI General-Purpose Agents Workflow Agents Personal Assistants RAG & Document AI Voice Agents

Trending

Also listed in

This tool also appears in

Data & Analytics › Data Labeling & Annotation Coding & Development › Open Source & Models

Explore

Browse AI tools by category

Content & Design Productivity & Business Coding & Development AI Agents & Automation Research & Education Wellness & Lifestyle Career Development Marketing & Growth Data & Analytics Customer Support & CX Finance E-commerce