About the DeepSeek R1 and V3 Models Used on Trae
TRAE IDE & SOLO information about the DeepSeek R1 and V3 models used on Trae.
What are DeepSeek-R1 and DeepSeek-V3?
DeepSeek-R1 and DeepSeek-V3 are advanced reasoning language models that are open source and can be downloaded and run independently from the DeepSeek website, supporting customization and deployment on different platforms. We do not rely on the DeepSeek API to generate outputs. Instead, we deploy these open source models on servers located in Southeast Asia. Both DeepSeek-R1 and DeepSeek-V3 are designed for code-related tasks, specifically for code Q&A and generation in chat and builder modes.
Will my data be sent to DeepSeek?
No data will be sent to DeepSeek. We use open source models and only run them within servers located in Southeast Asia. All Trae user data processing is conducted according to our privacy policy.
Do I have to use DeepSeek-R1 and DeepSeek-V3 on Trae?
The use of DeepSeek-R1 and DeepSeek-V3 is optional, and you can use any other models provided by Trae.
Model Features
DeepSeek-R1
- Type: Reasoning language model
- Specialty: Complex reasoning tasks
- Use Cases: Code problems requiring deep thinking and logical reasoning
- Advantages: Powerful problem analysis and solving capabilities
DeepSeek-V3
- Type: General language model
- Specialty: Code generation and understanding
- Use Cases: Daily programming tasks and code Q&A
- Advantages: Fast response and high-quality code generation
Usage Scenarios
Chat Mode
- Code problem solving
- Programming concept explanations
- Debugging suggestions
- Best practice guidance
Builder Mode
- Automatic code generation
- Project structure creation
- Feature implementation
- Code refactoring suggestions
Data Security and Privacy
Data Processing Location
- All data processing occurs on Southeast Asian servers
- No communication with DeepSeek official servers
- Complies with Trae privacy policy requirements
Privacy Protection Measures
- Data does not leave Trae's infrastructure
- Follows strict data protection standards
- User data is not used for model training
Model Selection Guide
When to Use DeepSeek-R1
- Complex algorithm design
- Architecture decision analysis
- Performance optimization problems
- Scenarios requiring deep reasoning
When to Use DeepSeek-V3
- Quick code generation
- Daily programming tasks
- Code completion
- Simple problem solving
When to Choose Other Models
- Specific domain professional needs
- Different language preferences
- Special performance requirements
- Personal usage habits
Technical Specifications
Deployment Architecture
- Server Location: Southeast Asia
- Deployment Method: Independent deployment, not API calls
- Model Version: Open source version
- Update Mechanism: Regular updates to the latest open source version
Performance Characteristics
- Response Speed: Optimized inference speed
- Accuracy: High-quality code generation
- Stability: 24/7 availability
- Scalability: Supports high concurrent usage
Frequently Asked Questions
Q: Will these models learn from my code?
A: No. We use pre-trained open source models and do not use user data for additional training.
Q: How can I switch to other models?
A: You can select other available models in Trae's model settings.
Q: How frequently are the models updated?
A: We regularly update to the latest open source versions released by DeepSeek.
Q: Do they support offline usage?
A: Currently, the models require an internet connection to use as they are deployed on our servers.
Feedback and Support
If you have privacy concerns about DeepSeek-R1 and DeepSeek-V3, please contact feedback@mail.trae.ai.
Other Support Methods
- Through the feedback feature within Trae IDE
- Visit our community forums
- Check related documentation and tutorials