GitHub Copilot: Custom Models & Enterprise AI Integration

GitHub Copilot is transforming into a gateway for custom AI models, moving beyond a single product. This shift is significant because developers care about which model is answering, data handling, inference costs, and enterprise security. GitHub moved Copilot to usage-based billing on June 1, 2026. This makes bringing in other model providers a crucial feature for budgeting and compliance. Copilot Business and Enterprise users can now use their own language model keys in Visual Studio Code. This includes providers like Anthropic, Gemini, OpenAI, OpenRouter, Azure, and local models via Ollama and Foundry Local. These models work in VS Code Chat but not for code completions. Usage for these custom models is billed by the chosen provider and doesn't count against Copilot quotas. This change addresses a major hurdle for enterprise adoption of AI coding tools. Companies can now route sensitive work to approved models or use their own volume discounts. The Copilot CLI also supports connecting to various endpoints, including local deployments, and offers an offline mode. This flexibility is key to scaling AI tools across organizations.