Support for bringing your own LLM models / custom model endpoints

[Omzig]

Enhancement Request

Summary

Add support for connecting GitHub Copilot CLI to custom/third-party LLM models (Google Cloud AI, Azure OpenAI, local models, etc.), similar to how Claude CLI allows users to specify a Google Cloud AI instance as a provider.

Motivation

Currently, GitHub Copilot CLI is tightly coupled to GitHub's default models. Unlike Claude CLI (which supports /model google-cloud-ai configuration), there's no built-in way to:

• Use enterprise-specific LLM deployments
• Connect to on-premises or private cloud models
• Switch between different model providers
• Use cost-optimized or specialized models

Benefits

• Enterprise flexibility
• Multi-cloud support
• Compatibility with private deployments
• Better cost management
• Parity with other CLI tools (Claude CLI)

Acceptance Criteria

• Users can configure a custom model endpoint
• Authentication credentials are securely stored
• CLI respects the custom model configuration
• Documentation is updated with examples