Use a Custom Model with GitHub Copilot

Stock models are fine. YOUR models are better. Let me show you end-to-end how to surface a custom LLM from Azure AI Foundry in GitHub Copilot. === TIMESTAMPS === 00:00 - What we're building today 00:45 - BYOK is in public preview (caveats apply) 01:15 - Creating your Foundry resource in Azure 02:30 - Projects, Key Vault, App Insights integration 03:15 - Private endpoints and compliance controls 04:00 - Deploying GPT-5 in the Foundry portal 04:45 - Grabbing your endpoint URI and API key 05:30 - Fine-tuning and agent capabilities (teaser) 06:00 - GitHub Enterprise Cloud admin setup 06:45 - Adding your custom model via BYOK 08:00 - Org access and display name config 09:00 - Token limits and cost governance 09:45 - VS Code extensions you need 11:00 - Selecting your model in the Copilot picker 11:45 - Testing with a real prompt === KEY RESOURCES === GitHub BYOK Announcement: https://github.blog/changelog/2025-11-20-enterprise-bring-your-own-key-byok-for-github-copilot-is-now-in-public-preview/ VS Code Language Models Docs: https://code.visualstudio.com/docs/copilot/customization/language-models Azure AI Foundry Portal: https://ai.azure.com GitHub Copilot Custom Models Docs: https://docs.github.com/copilot === KEY TAKEAWAYS === → BYOK lets you connect Anthropic, Microsoft Foundry, OpenAI, or xAI models to Copilot → Usage bills through your provider - doesn't count against Copilot quotas → Requires GitHub Enterprise Cloud + Copilot Enterprise licenses → Enterprise admins control which orgs can access custom models → Watch your token limits - you're paying for inputs AND outputs === ABOUT ME === Tim Warner | Microsoft MVP | 28+ years in Microsoft tech Principal Staff Author at Pluralsight https://TechTrainerTim.com #GitHubCopilot #AzureAI #BYOK #CustomModels #AIFoundry #VSCode #CopilotEnterprise #Microsoft