Skip to content
AI Radar by Wagner Solution
CS
Menu
context Tools and apps

GitHub is shutting down Spark and Models, directing developers to GitHub Copilot

only one source so far

As of 4 August 2026, GitHub stopped new user registrations and the creation of new applications in Spark; existing applications can be exported until 31 August 2026. The GitHub Models inference service for the llm() function shut down on 30 July 2026; applications that do not use it continue to work.

As of 4 August 2026, GitHub stopped accepting new users and allowing new applications to be created in GitHub Spark. Existing users can still access Spark until 31 August 2026 to export applications they have already created. According to the company, applications deployed earlier will continue to work after Spark shuts down.

Separately, GitHub Models, which Spark used as the inference backend for the llm() function, shut down on 30 July 2026. From that date, llm() calls stopped working. Applications that do not use the llm() function are not affected by the shutdown of GitHub Models. If an application uses llm(), the company says the user must replace the inference service with their own provider, supply their own API key and manage the associated payments themselves, because GitHub no longer provides the model or tokens.

GitHub explains the shutdown by saying that AI models and tools for agentic development have advanced significantly since Spark launched, and developers increasingly choose integrated workflows directly in GitHub Copilot, specifically in VS Code, Copilot CLI and the GitHub Copilot app. The company therefore says it is adapting its offering to these changes.

What changed

Why it matters

Developers using GitHub Spark have a limited window to export their applications and must urgently arrange an alternative source of inference for those that depend on llm(); otherwise, their AI features will stop working and they will incur new direct API costs.

Two audiences, two different impacts

What this means

01

For individuals

Anyone with applications created in GitHub Spark must export them by 31 August 2026 and, if they use the llm() function, arrange their own API key and model provider; otherwise, the AI features will stop working.

What to do Open your applications in Spark, look for llm() calls in the code and, if any are found, replace them with inference from another provider; export applications from Spark by 31 August 2026.
More practical updates →
02

For a business

Companies that used GitHub Spark for rapid prototyping must move their workflows to GitHub Copilot and arrange and pay for an alternative inference provider themselves for applications that depend on GitHub Models.

Development
What to decide Map internal applications built on GitHub Spark, arrange an alternative inference provider and a budget for API keys for those that call llm(), and plan the transition of development workflows to GitHub Copilot (VS Code…
More business impacts →
AI inference deprecation GitHub Copilot GitHub Models GitHub Spark migration

Check the original

Event sources

only one source so far · 1 publisher, 0 independent. We count feeds from the same owner only once.

1
GitHub Changelog (Copilot and AI features) primary source · first detected Upcoming deprecation of GitHub Spark on github.com