HuggingFace | Paper | GitHub | HomePage

# Introduction OpenMobile-2-9B is a mobile GUI agent fine-tuned from [Qwen3.5-9B](https://huggingface.co/Qwen/Qwen3.5-9B) on [OpenMobile-Data-v2](https://huggingface.co/datasets/OpenMobile-2/OpenMobile-Data-v2). It completes Android tasks from screenshots, and can call app-native tools when the foreground app exposes them. ## Results | Model | AndroidWorld Pass@1 | MobileWorld GUI | MobileGym SR | MobileGym++ Hybrid | SPA-Bench L3 | |---|---:|---:|---:|---:|---:| | Qwen3.5-9B | 54.3 | 12.8 | 9.4 | 12.1 | 25.5 | | OpenMobile-1-8B | 64.7 | 17.7 | 13.7 | 7.0 | -- | | UI-Venus-1.5-8B | 73.7 | -- | 15.4 | 16.3 | 29.8 | | GUI-Owl-1.5-8B | 69.0 | 38.2 | 17.2 | 12.6 | -- | | MAI-UI-8B | 70.7 | 27.5 | 16.8 | 18.6 | -- | | **OpenMobile-2-9B** | **74.0** | **38.5** | **38.3** | **46.5** | **48.9** | ## Deploy The commands below use a context length of 65536 and tensor parallel 2 **vLLM** We use vLLM 0.19.1. ```bash pip install vllm==0.19.1 ``` ```bash vllm serve OpenMobile-2/OpenMobile-2-9B \ --tensor-parallel-size 2 \ --max-model-len 65536 \ --gpu-memory-utilization 0.90 \ --max-num-seqs 16 \ --reasoning-parser qwen3 \ --enable-auto-tool-choice \ --tool-call-parser qwen3_coder \ --gdn-prefill-backend triton ``` **SGLang** We use SGLang 0.5.10.post1. ```bash pip install sglang==0.5.10.post1 ``` ```bash python -m sglang.launch_server \ --model-path OpenMobile-2/OpenMobile-2-9B \ --tp-size 2 \ --mem-fraction-static 0.8 \ --context-length 65536 \ --reasoning-parser qwen3 \ --tool-call-parser qwen3_coder \ --enable-multimodal ``` ## Citation If you find this model useful, please cite our paper: ```bibtex @article{openmobile2, title={OpenMobile-2: Building Versatile Mobile Agents with Scalable Environments and App-Native Tools}, author={Chenyang Yan and Yingying Zhang and Kanzhi Cheng and Qiushi Sun and Zhengyuan Pan and Zheng Ma and Hang Yan and Yian Wang and Nuo Chen and Jialin Cao and Xingdong Gong and Wenpo Song and Han Chen and Zichen Ding and Fangzhi Xu and Shujian Huang and Xinyu Dai and Yichen Liu and Tiankuo Yao and Bo Wang and Ben Kao and Jianbing Zhang and Lewei Lu and Dahua Lin}, journal={arXiv preprint arXiv:2610.14157}, year={2026}, url={https://arxiv.org/abs/2610.14157} } ```

Full Model Card (README)