Available Models
AI models available in Playground
Playground offers multiple AI models optimized for computer use tasks. Each model has different capabilities for UI understanding, visual processing, and reasoning.
Selecting a Model
Use the model dropdown at the top of the chat panel to choose a model. The selection applies to new messages only - previous messages keep their original model.
Choose different models based on:
- Task complexity
- Required capabilities (UI interaction, visual analysis, reasoning)
- Speed requirements
Models
Doubao 1.5 UI-TARS 250428
Advanced UI understanding with precise element interaction capabilities. Shows reasoning process before taking actions.
Best for:
- Complex UI navigation
- Multi-step workflows
- Tasks requiring precise clicking
Example tasks:
Navigate to GitHub and find the Lybic repositoryOpen Chrome, search for "Python tutorials", and compare top 3 resultsDoubao 1.5 Vision Pro
Enhanced visual processing and screen understanding. Optimized for analyzing visual content.
Best for:
- Visual analysis
- Content comprehension
- Image-heavy workflows
Example tasks:
Analyze this webpage and summarize key informationCompare product images on this e-commerce siteDoubao Seed 1.6 Vision
Latest generation model with advanced reasoning capabilities. Good general-purpose choice.
Best for:
- Complex reasoning tasks
- Multi-step operations
- Latest model features
Example tasks:
Research quantum computing basics and summarizePlan and execute a multi-step installationDoubao Seed 1.6 Flash
Optimized for speed and efficiency. Good for simple, frequent tasks.
Best for:
- Quick operations
- Simple tasks
- Fast response needs
Example tasks:
Open Chrome and go to lybic.cnCreate a text file with specific contentModel Comparison
| Feature | UI-TARS | Vision Pro | Seed 1.6 Vision | Seed 1.6 Flash |
|---|---|---|---|---|
| UI Interaction | Excellent | Good | Moderate | Moderate |
| Visual Analysis | Good | Excellent | Good | Moderate |
| Reasoning | Good | Good | Excellent | Moderate |
| Speed | Moderate | Moderate | Moderate | Fast |
| Thinking Process | Yes | Yes | Yes | Yes |
Selection Guide
For UI-Heavy Tasks
Use Doubao 1.5 UI-TARS for tasks requiring precise UI interaction and element clicking.
For Visual Content
Use Doubao 1.5 Vision Pro when analyzing images, charts, or visual information is important.
For Complex Reasoning
Use Doubao Seed 1.6 Vision for tasks requiring multi-step planning and decision-making.
For Simple Tasks
Use Doubao Seed 1.6 Flash for quick, straightforward operations where speed matters.
Thinking Process
All models show their reasoning process before taking actions. You'll see the AI's step-by-step plan in chat before it executes.
This helps you:
- Understand the AI's approach
- Catch issues before execution
- Learn task breakdown patterns
Resource Usage
All models consume Agent Credits from your organization's package. Credit usage scales with:
- Task complexity
- Conversation length
- Model capabilities used
Monitor usage in your Dashboard billing section.
Tips
Be Specific with UI Elements
When using UI-focused models, describe elements clearly:
Click the blue 'Sign In' button in the top right cornerTry Different Models
For challenging tasks, try different models to compare results. Change the model in a new chat and give the same instruction.
Start Simple
Begin with straightforward tasks to understand each model's behavior before attempting complex operations.