Copilot Models
Unlock third-party large language model extensions for GitHub Copilot.
Seamlessly integrate mainstream LLMs like DeepSeek and Zhipu AI.
One-click switching and native panel compatibility.
Features
- Multi-Model Support: Supports DeepSeek V4 series and Zhipu AI GLM-5 series
- Tool Calling: Supports Copilot Chat tool calling functionality
- Thinking Mode: Supports model thinking/reasoning mode
- Secure Authentication: API keys securely stored in VS Code SecretStorage
- Log Debugging: Complete logging system for troubleshooting
Documentation
This project supports multiple languages for documentation:
Quick Start
1. Install Extension
Install the "Copilot Models" extension from the VS Code Extension Marketplace.
Method 1: Command Palette
- Press
Ctrl+Shift+P (macOS: Cmd+Shift+P)
- Type
Copilot Models: Set API Key
- Press Enter, then enter the provider name (e.g.,
deepseek or bigmodel)
- Enter the corresponding API key
DeepSeek API Key can be obtained from DeepSeek Platform
Zhipu AI API Key can be obtained from Zhipu AI Open Platform
Method 2: Settings Page
- Press
Ctrl+, to open VS Code settings
- Search for
copilot-models
- Configure provider enablement, base URLs, and debug options
Note: API keys are managed through the command palette.
They are stored in VS Code SecretStorage, not as plain settings.
Tips:
3. Start Using
- Open GitHub Copilot Chat panel
- Click on the model selector
- Select the model to use
- Start chatting
Supported Models
DeepSeek
| Model |
Description |
Tool Calling |
Thinking Mode |
| DeepSeek V4 Flash |
Fast response, supports tool calling |
✅ |
✅ |
| DeepSeek V4 Pro |
Deep thinking, stronger reasoning |
❌ |
✅ |
Zhipu AI (BigModel)
| Model |
Context |
Output |
Tool Calling |
Thinking Mode |
| GLM-5.1 |
200K |
128K |
✅ |
✅ |
| GLM-5-Turbo |
200K |
128K |
✅ |
✅ |
| GLM-5 |
200K |
128K |
✅ |
✅ |
Configuration Options
The following configuration options are available in VS Code settings:
copilot-models.deepseek.enabled: Enable the DeepSeek provider.
Default: true.
copilot-models.deepseekBaseUrl: DeepSeek API base URL.
Default: https://api.deepseek.com.
copilot-models.bigmodel.enabled: Enable the BigModel provider.
Default: true.
copilot-models.bigmodelBaseUrl: BigModel API base URL.
Default: https://open.bigmodel.cn/api/paas/v4.
copilot-models.modelIdOverrides: Override model IDs for custom endpoints.
Default: {}.
copilot-models.maxTokens: Maximum generated tokens.
Default: 0 (unlimited).
copilot-models.timeoutMs: API request timeout in milliseconds.
Default: 60000.
copilot-models.maxRetries: Maximum number of API request retries.
Default: 1.
copilot-models.debugMode: Log level.
Default: minimal.
Note: API keys are managed through the command palette.
They are stored in VS Code SecretStorage, not as plain settings.
Commands
| Command |
Description |
Copilot Models: Set API Key |
Configure API key (select provider first) |
Copilot Models: Clear API Key |
Clear API key (select provider first) |
Copilot Models: Open Settings |
Open extension settings |
Copilot Models: Show Log |
Show log panel |
Copilot Models: Clear Log |
Clear logs |
Copilot Models: Refresh Models |
Refresh model list |
Debugging
If you encounter issues, you can check the logs:
- Press
Ctrl+Shift+P, type Copilot Models: Show Log
- Logs will be displayed in the "Copilot Models" output panel
- For more detailed logs, change
Debug Mode to verbose in settings
License
Apache-2.0
| |