GitHub Models API retired: Open WebUI 404 and migration.

Why Open WebUI may show a GitHub Models 404

A February 2026 Open WebUI issue #21257 recorded model refresh requests to https://models.github.ai/inference/models returning 404 while GitHub Models was active. That report is historical. After GitHub’s retirement, an error from models.github.ai is not repaired by the old catalog-path workaround.

The currently published Open WebUI provider guide still lists GitHub Models as a provider without the standard /models endpoint and suggests a manual model-ID allowlist. That advice cannot make a retired upstream available. A saved or allow-listed model name is only a local selector entry; it does not create an API service.

Move the connection to an active provider

1. Disable or remove the retired GitHub Models connection in Open WebUI. Do not reuse its GitHub token as a key for another provider.
2. Select a provider or gateway whose current documentation explicitly supports the Chat Completions API used by your client.
3. In Settings → Admin → Connections, enter that provider’s documented API base URL and a key issued by that provider. Match the base path, authentication method and region exactly.
4. Use the provider’s current model ID. If its API does not expose /models, Open WebUI allows you to add documented model IDs manually—but only for an active endpoint.
5. Save the connection, then test a short chat with a model you are authorized to use. Model-list discovery and successful authentication do not prove that inference works.

For a live OpenAI-compatible connection, follow the Open WebUI setup and request-location checks. To compare documented routes before choosing one, see the API route comparison.

Choose the GitHub product that matches the task

• Need a general model API for an application? GitHub’s retirement notice points new and existing model-access projects to Azure AI Foundry. Check its own current access, pricing and API instructions.
• Need AI inside GitHub workflows? Review GitHub Copilot’s current product and client integrations. Copilot is not a drop-in replacement base URL for an OpenAI-compatible API connection.

Compatibility is defined by the active provider’s API behavior, not by a model name appearing in a client’s selector. Confirm the supported endpoint and features with the provider before moving production traffic.

Sources and scope

• GitHub Models retirement notice — states the July 30, 2026 retirement and directs users to Azure AI Foundry or Copilot.
• Open WebUI OpenAI-compatible provider guide — connection setup, model-ID allowlists and provider verification behavior; its GitHub Models row is now superseded by the retirement notice.
• Open WebUI issue #21257 — a dated report of the former /inference/models 404, not evidence of current service availability.

Checked 2 November 2026 against GitHub’s current retirement notice and Open WebUI’s current provider guide. This is a documentation-based diagnosis; no GitHub, Azure, Open WebUI, provider or model API request was made.

Free API access and limits

Check the site free API access and limits before trying to obtain a site key. Read the current status and request limits before using a site key. It is not a source of official provider keys.

ChinaGenAI is not affiliated with the providers named here. This release contains setup guidance and explicitly labeled fictional examples, not claimed live model benchmarks.