---
license: apache-2.0
language:
- en
pipeline_tag: image-text-to-text
tags:
- multimodal
- gui
library_name: transformers
---

# UI-TARS-2B-SFT
[UI-TARS-2B-SFT](https://huggingface.co/bytedance-research/UI-TARS-2B-SFT)  |  
[UI-TARS-2B-gguf](https://huggingface.co/bytedance-research/UI-TARS-2B-gguf)  |  
[UI-TARS-7B-SFT](https://huggingface.co/bytedance-research/UI-TARS-7B-SFT)  | 
[**UI-TARS-7B-DPO**](https://huggingface.co/bytedance-research/UI-TARS-7B-DPO)(Recommended)  | 
[UI-TARS-7B-gguf](https://huggingface.co/bytedance-research/UI-TARS-7B-gguf)  |  
[UI-TARS-72B-SFT](https://huggingface.co/bytedance-research/UI-TARS-72B-SFT)  | 
[**UI-TARS-72B-DPO**](https://huggingface.co/bytedance-research/UI-TARS-72B-DPO)(Recommended)
## Introduction

UI-TARS is a next-generation native GUI agent model designed to interact seamlessly with graphical user interfaces (GUIs) using human-like perception, reasoning, and action capabilities. Unlike traditional modular frameworks, UI-TARS integrates all key components—perception, reasoning, grounding, and memory—within a single vision-language model (VLM), enabling end-to-end task automation without predefined workflows or manual rules.
<!-- ![Local Image](figures/UI-TARS.png) -->
<p align="center">
    <img src="https://github.com/bytedance/UI-TARS/blob/main/figures/UI-TARS-vs-Previous-SOTA.png?raw=true" width="90%"/>
<p>
<p align="center">
    <img src="https://github.com/bytedance/UI-TARS/blob/main/figures/UI-TARS.png?raw=true" width="90%"/>
<p>

<!-- ![Local Image](figures/UI-TARS-vs-Previous-SOTA.png) -->

This repository contains the model for the paper [UI-TARS: Pioneering Automated GUI Interaction with Native Agents](https://huggingface.co/papers/2501.12326).

Code: https://github.com/bytedance/UI-TARS

## Performance
**Perception Capabilty Evaluation**
| Model                     | VisualWebBench | WebSRC  | SQAshort |
|---------------------------|---------------|---------|----------|
| Qwen2-VL-7B              | 73.3          | 81.8    | 84.9     |
| Qwen-VL-Max              | 74.1          | 91.1    | 78.6     |
| Gemini-1.5-Pro           | 75.4          | 88.9    | 82.2     |
| UIX-Qwen2-7B             | 75.9          | 82.9    | 78.8     |
| Claude-3.5-Sonnet        | 78.2          | 90.4    | 83.1     |
| GPT-4o                   | 78.5          | 87.7    | 82.3     |
| **UI-TARS-2B**          | 72.9          | 89.2    | 86.4     |
| **UI-TARS-7B**          | 79.7          | **93.6** | 87.7     |
| **UI-TARS-72B**         | **82.8**      | 89.3    | **88.6** |

**Grounding Capability Evaluation**
- **ScreenSpot Pro**

| Agent Model              | Dev-Text | Dev-Icon | Dev-Avg | Creative-Text | Creative-Icon | Creative-Avg | CAD-Text | CAD-Icon | CAD-Avg | Scientific-Text | Scientific-Icon | Scientific-Avg | Office-Text | Office-Icon | Office-Avg | OS-Text | OS-Icon | OS-Avg | Avg-Text | Avg-Icon | Avg |
|--------------------------|----------|----------|----------|--------------|--------------|--------------|---------|---------|---------|---------------|---------------|---------------|------------|------------|------------|--------|--------|--------|---------|---------|------|
| QwenVL-7B               | 0.0      | 0.0      | 0.0      | 0.0          | 0.0          | 0.0          | 0.0     | 0.0     | 0.0     | 0.7           | 0.0           | 0.4           | 0.0        | 0.0        | 0.0        | 0.0    | 0.0    | 0.0    | 0.1     | 0.0     | **0.1**  |
| GPT-4o                  | 1.3      | 0.0      | 0.7      | 1.0          | 0.0          | 0.6          | 2.0     | 0.0     | 1.5     | 2.1           | 0.0           | 1.2           | 1.1        | 0.0        | 0.9        | 0.0    | 0.0    | 0.0    | 1.3     | 0.0     | **0.8**  |
| SeeClick                | 0.6      | 0.0      | 0.3      | 1.0          | 0.0          | 0.6          | 2.5     | 0.0     | 1.9     | 3.5           | 0.0           | 2.0           | 1.1        | 0.0        | 0.9        | 2.8    | 0.0    | 1.5    | 1.8     | 0.0     | **1.1**  |
| Qwen2-VL-7B             | 2.6      | 0.0      | 1.3      | 1.5          | 0.0          | 0.9          | 0.5     | 0.0     | 0.4     | 6.3           | 0.0           | 3.5           | 3.4        | 1.9        | 3.0        | 0.9    | 0.0    | 0.5    | 2.5     | 0.2     | **1.6**  |
| OS-Atlas-4B            | 7.1      | 0.0      | 3.7      | 3.0          | 1.4          | 2.3          | 2.0     | 0.0     | 1.5     | 9.0           | 5.5           | 7.5           | 5.1        | 3.8        | 4.8        | 5.6    | 0.0    | 3.1    | 5.0     | 1.7     | **3.7**  |
| ShowUI-2B              | 16.9     | 1.4      | 9.4      | 9.1          | 0.0          | 5.3          | 2.5     | 0.0     | 1.9     | 13.2          | 7.3           | 10.6          | 15.3       | 7.5        | 13.5       | 10.3   | 2.2    | 6.6    | 10.8    | 2.6     | **7.7**  |
| CogAgent-18B           | 14.9     | 0.7      | 8.0      | 9.6          | 0.0          | 5.6          | 7.1     | 3.1     | 6.1     | 22.2          | 1.8           | 13.4          | 13.0       | 0.0        | 10.0       | 5.6    | 0.0    | 3.1    | 12.0    | 0.8     | **7.7**  |
| Aria-UI                | 16.2     | 0.0      | 8.4      | 23.7         | 2.1          | 14.7         | 7.6     | 1.6     | 6.1     | 27.1          | 6.4           | 18.1          | 20.3       | 1.9        | 16.1       | 4.7    | 0.0    | 2.6    | 17.1    | 2.0     | **11.3**  |
| UGround-7B             | 26.6     | 2.1      | 14.7     | 27.3         | 2.8          | 17.0         | 14.2    | 1.6     | 11.1    | 31.9          | 2.7           | 19.3          | 31.6       | 11.3       | 27.0       | 17.8   | 0.0    | 9.7    | 25.0    | 2.8     | **16.5**  |
| Claude Computer Use      | 22.0  | 3.9   | 12.6  | 25.9  | 3.4   | 16.8  | 14.5  | 3.7   | 11.9  | 33.9  | 15.8  | 25.8  | 30.1  | 16.3  | 26.9  | 11.0  | 4.5   | 8.1   | 23.4  | 7.1  | **17.1**  |
| OS-Atlas-7B              | 33.1  | 1.4   | 17.7  | 28.8  | 2.8   | 17.9  | 12.2  | 4.7   | 10.3  | 37.5  | 7.3   | 24.4  | 33.9  | 5.7   | 27.4  | 27.1  | 4.5   | 16.8  | 28.1  | 4.0  | **18.9**  |
| UGround-V1-7B            | -     | -     | 35.5  | -     | -     | 27.8  | -     | -     | 13.5  | -     | -     | 38.8  | -     | -     | 48.8  | -     | -     | 26.1  | -     | -    | **31.1**  |
| **UI-TARS-2B**        | 47.4     | 4.1      | 26.4     | 42.9         | 6.3          | 27.6         | 17.8    | 4.7     | 14.6    | 56.9          | 17.3          | 39.8          | 50.3       | 17.0       | 42.6       | 21.5   | 5.6    | 14.3   | 39.6    | 8.4     | **27.7**  |
| **UI-TARS-7B**        | 58.4     | 12.4     | 36.1     | 50.0         | 9.1          | 32.8         | **20.8**| 9.4     | **18.0**| 63.9          | **31.8**      | **50.0**      | **63.3**   | 20.8       | 53.5       | 30.8   | **16.9**| 24.5   | 47.8    | 16.2    | **35.7**  |
| **UI-TARS-72B**       | **63.0** | **17.3** | **40.8** | **57.1**     | **15.4**     | **39.6**     | 18.8    | **12.5**| 17.2    | **64.6**      | 20.9          | 45.7          | **63.3**   | **26.4**   | **54.8**   | **42.1**| 15.7    | **30.1**| **50.9**| **17.5**| **38.1**  |


- **ScreenSpot v2**

| Method |  Mobile-Text | Mobile-Icon/Widget | Desktop-Text | Desktop-Icon/Widget | Web-Text | Web-Icon/Widget | Avg |
|--------|-------------|-------------|-------------|-------------|-------------|---------|---------|
| **Agent Framework**  | | | | | | | |
| GPT-4o (SeeClick)  | 85.2 | 58.8 | 79.9 | 37.1 | 72.7 | 30.1 | **63.6** |
| GPT-4o (OS-Atlas-4B)  | 95.5 | 75.8 | 79.4 | 49.3 | 90.2 | 66.5 | **79.1** |
| GPT-4o (OS-Atlas-7B)  | 96.2 | 83.4 | 89.7 | 69.3 | **94.0** | 79.8 | **87.1** |
| **Agent Model**  | | | | | | | |
| SeeClick  | 78.4 | 50.7 | 70.1 | 29.3 | 55.2 | 32.5 | **55.1** |
| OS-Atlas-4B  | 87.2 | 59.7 | 72.7 | 46.4 | 85.9 | 63.1 | **71.9** |
| OS-Atlas-7B  | 95.2 | 75.8 | 90.7 | 63.6 | 90.6 | 77.3 | **84.1** |
| **Our Model**  | | | | | | | |
| **UI-TARS-2B**  | 95.2 | 79.1 | 90.7 | 68.6 | 87.2 | 78.3 | **84.7** |
| **UI-TARS-7B** | **96.9** | **89.1** | **95.4** | 85.0 | 93.6 | 85.2 | **91.6** |
| **UI-TARS-72B**  | 94.8 | 86.3 | 91.2 | **87.9** | 91.5 | **87.7** | **90.3** |


**Online Agent Capability Evaluation**

| Method |  OSWorld (Online) | AndroidWorld (Online) |
|--------|-------------------|------------------|
| **Agent Framework**  | | |
| GPT-4o (UGround)  | - | 32.8 |
| GPT-4o (Aria-UI)  | 15.2 | 44.8 |
| GPT-4o (Aguvis-7B)  | 14.8 | 37.1 |
| GPT-4o (Aguvis-72B)  | 17.0 | - |
| GPT-4o (OS-Atlas-7B)  | 14.6 | - |
| **Agent Model**  | | |
| GPT-4o  | 5.0 | 34.5 (SoM) |
| Gemini-Pro-1.5  | 5.4 | 22.8 (SoM) |
| Aguvis-72B  | 10.3 | 26.1 |
| Claude Computer-Use  | 14.9 (15 steps) | 27.9 |
| Claude Computer-Use  | 22.0 (50 steps) | - |
| **Our Model**  | | |
| **UI-TARS-7B-SFT**  | 17.7 (15 steps) | 33.0 |
| **UI-TARS-7B-DPO**  | 18.7 (15 steps) | - |
| **UI-TARS-72B-SFT**  | 18.8 (15 steps) | **46.6** |
| **UI-TARS-72B-DPO**  | **22.7** (15 steps) | - |
| **UI-TARS-72B-DPO**  | **24.6** (50 steps) | - |


## How to use
The usage will be the same as LLaVA with a few additional tokens.
Please refer to https://github.com/haotian-liu/LLaVA

## License
UI-TARS is licensed under the Apache License 2.0.