Scale computer-use 2.0 with open-source drivers, cross-OS sandboxes, and benchmarks for training, evaluation, and data generation.
Rust agentai-agentapplecomputer-usecomputer-use-agentcontainerizationcuadesktop-automationhacktoberfestlumemacosmanusoperatorswiftvirtualizationvirtualization-frameworkwindowswindows-sandbox
document cloud vlm models
Sarina Li3b2722ad8ab9
diff --git a/README.md b/README.mdindex 4b46c9f161..57c0d0d098 100644--- a/README.md+++ b/README.md@@ -219,21 +219,22 @@ These are the valid model configurations for `ComputerAgent(model="...")`:The following table shows which capabilities are supported by each model:-| Model | Computer-Use | Grounding | Tools | VLM |-| -------------------------------------------------------------------------------------------------------------------------------- | :----------: | :-------: | :---: | :-: |-| [Claude Sonnet/Haiku](https://docs.claude.com/en/docs/agents-and-tools/tool-use/computer-use-tool#how-to-implement-computer-use) | π₯οΈ | π― | π οΈ | ποΈ |-| [OpenAI CU Preview](https://platform.openai.com/docs/models/computer-use-preview) | π₯οΈ | π― | | ποΈ |-| [Qwen3 VL](https://huggingface.co/collections/Qwen/qwen3-vl) | π₯οΈ | π― | π οΈ | ποΈ |-| [GLM-V](https://huggingface.co/THUDM/glm-4v-9b) | π₯οΈ | π― | π οΈ | ποΈ |-| [Gemini CU Preview](https://ai.google.dev/gemini-api/docs/computer-use) | π₯οΈ | π― | | ποΈ |-| [InternVL](https://huggingface.co/OpenGVLab/InternVL3_5-1B) | π₯οΈ | π― | π οΈ | ποΈ |-| [UI-TARS](https://huggingface.co/ByteDance-Seed/UI-TARS-1.5-7B) | π₯οΈ | π― | π οΈ | ποΈ |-| [UI-TARS-2](https://cua.ai/dashboard/vlm-router) | π₯οΈ | π― | π οΈ | ποΈ |-| [OpenCUA](https://huggingface.co/xlangai/OpenCUA-7B) | | π― | | |-| [GTA](https://huggingface.co/HelloKKMe/GTA1-7B) | | π― | | |-| [Holo](https://huggingface.co/Hcompany/Holo1.5-3B) | | π― | | |-| [Moondream](https://huggingface.co/moondream/moondream3-preview) | | π― | | |-| [OmniParser](https://github.com/microsoft/OmniParser) | | π― | | |+| Model | Computer-Use | Grounding | Tools | VLM | Cloud |+| -------------------------------------------------------------------------------------------------------------------------------- | :----------: | :-------: | :---: | :-: | :---: |+| [Claude Sonnet/Haiku](https://docs.claude.com/en/docs/agents-and-tools/tool-use/computer-use-tool#how-to-implement-computer-use) | π₯οΈ | π― | π οΈ | ποΈ | βοΈ |+| [Claude Opus](https://docs.claude.com/en/docs/agents-and-tools/tool-use/computer-use-tool#how-to-implement-computer-use) | π₯οΈ | π― | π οΈ | ποΈ | βοΈ |+| [OpenAI CU Preview](https://platform.openai.com/docs/models/computer-use-preview) | π₯οΈ | π― | | ποΈ | |+| [Qwen3 VL](https://huggingface.co/collections/Qwen/qwen3-vl) | π₯οΈ | π― | π οΈ | ποΈ | βοΈ |+| [GLM-V](https://huggingface.co/THUDM/glm-4v-9b) | π₯οΈ | π― | π οΈ | ποΈ | |+| [Gemini CU Preview](https://ai.google.dev/gemini-api/docs/computer-use) | π₯οΈ | π― | | ποΈ | |+| [InternVL](https://huggingface.co/OpenGVLab/InternVL3_5-1B) | π₯οΈ | π― | π οΈ | ποΈ | |+| [UI-TARS](https://huggingface.co/ByteDance-Seed/UI-TARS-1.5-7B) | π₯οΈ | π― | π οΈ | ποΈ | |+| [UI-TARS-2](https://cua.ai/dashboard/vlm-router) | π₯οΈ | π― | π οΈ | ποΈ | βοΈ |+| [OpenCUA](https://huggingface.co/xlangai/OpenCUA-7B) | | π― | | | |+| [GTA](https://huggingface.co/HelloKKMe/GTA1-7B) | | π― | | | |+| [Holo](https://huggingface.co/Hcompany/Holo1.5-3B) | | π― | | | |+| [Moondream](https://huggingface.co/moondream/moondream3-preview) | | π― | | | |+| [OmniParser](https://github.com/microsoft/OmniParser) | | π― | | | |**Legend:**@@ -241,6 +242,7 @@ The following table shows which capabilities are supported by each model:- π― **Grounding**: UI element detection and click coordinate prediction- π οΈ **Tools**: Support for function calling beyond screen interaction- ποΈ **VLM**: Vision-language understanding+- βοΈ **Cloud**: Supported on Cua VLM**Composition Examples:**@@ -381,6 +383,20 @@ Learn more in the [SOM documentation](./libs/python/som/README.md).## 2025+### December 2025++- **Cloud VLM Platform**: Support for Claude Opus, Qwen3 VL 235B, and UI-TARS-2 on Cua VLM cloud infrastructure+- **QEMU Container Support**: Native Linux and Windows container execution via QEMU virtualization++### November 2025++- **Generic VLM Provider**: Expanded support for custom VLM providers and model configurations+- **NeurIPS 2025**: Coverage of computer-use agent research papers and developments ([Blog Post](https://cua.ai/blog/neurips-2025-cua-papers))++### October 2025++- **Agent SDK Improvements**: Enhanced model support and configuration options+### September 2025- **Hack the North Competition**: First benchmark-driven hackathon track with guaranteed YC interview prize. Winner achieved 68.3% on OSWorld-Tiny ([Blog Post](https://www.cua.ai/blog/hack-the-north))