---
title: Model management
seoTitle: "Langfuse AI Gateway model management (beta)"
description: See which models the Langfuse AI Gateway can call, enable or disable models per provider credential, and choose the right model ID for your client.
---

# Model management

  **Beta.** The Langfuse AI Gateway is in public beta and can still change.

As an admin, you can manage which models are allowed to be used. This page explains where the model list comes from, how admins choose models, and which model ID to send.

  ![Provider credential page with a searchable model list, an Enabled switch per model, and Enable all and Disable all buttons](/images/docs/ai-gateway/models-manage.png)

## Available models [#available-models]

You can find a list of every model gateway keys can call under **Organization settings → AI Gateway → Models**. Everyone with gateway access can open this page, including Viewers and the AI Gateway role.

The list comes from each provider credential: the gateway pings the provider's own `/models` endpoint to check which models the credential can access.

  If a provider rejects a credential during model discovery, for example
  because the key was rotated, the credential is marked as failing and the
  gateway stops routing to it. Update the key or select **Retry** in
  **Organization settings → AI Gateway → Setup** to restore it.

## Manage models [#manage-models]

Gateway managers (Owners and Admins) can configure models per provider credential:

1. Go to **Organization settings → AI Gateway → Setup** and open a provider credential.
2. Turn models on or off with the **Enabled** switch.
3. Select **Refresh provider models** to fetch the provider's current list.

A credential with every model enabled also allows models that the provider publishes later. As soon as you turn one model off, the credential keeps an explicit list. Select **Enable all** to go back to allowing every model.

## Model IDs [#model-ids]

Model IDs have the form `<credential>:<model>`, for example `openai-prod:<model>`, so the gateway knows which credential to use. For coding agents that send the provider's own model names, IDs without a prefix also work if they start with `gpt-`, `codex-`, `ft:`, or `claude-`, and the gateway uses the oldest matching credential. Renaming a credential changes its prefix, so update clients that use the old IDs.

## Troubleshooting and FAQ [#troubleshooting-and-faq]

### Why can't I call a new model? [#new-models]

A model the provider just released is available right away only on credentials that have every model enabled. If an admin turned off any model on a credential, the credential keeps an explicit list, and new models stay off until an admin enables them under **Organization settings → AI Gateway → Setup**. Select **Refresh provider models** there if the model doesn't appear in the list yet.

[Ask another question](/docs/ask-ai)

<!-- agent-instructions -->

---

## Agent Instructions

This page is part of the [Langfuse](https://langfuse.com) documentation, published as plain Markdown for AI agents. Every page is available as Markdown by appending `.md` to its URL, or by sending an `Accept: text/markdown` header. This page: `https://langfuse.com/docs/gateway/features/models.md`.

### Querying these docs

If the answer is not on this page, query the documentation instead of guessing:

- **Semantic search** across all Langfuse docs, returning an answer with the relevant pages and excerpts. Ask a specific, self-contained question:

  ```bash
  curl -sG "https://langfuse.com/api/search-docs" --data-urlencode "query=How do I trace a LangGraph agent?"
  ```

- **Index of every page**: <https://langfuse.com/llms.txt>, with per-section indexes [llms-docs.txt](https://langfuse.com/llms-docs.txt), [llms-integrations.txt](https://langfuse.com/llms-integrations.txt), and [llms-self-hosting.txt](https://langfuse.com/llms-self-hosting.txt).

### Before writing Langfuse code

- **Install the [Langfuse Agent Skill](https://langfuse.com/docs/api-and-data-platform/features/agent-skill).** It encodes Langfuse's own best practices for instrumentation, prompt management, and evaluation, and materially improves results.
- **Read [What does a good trace look like?](https://langfuse.com/docs/observability/best-practices.md)** before instrumenting an application.
- **Verify endpoints, parameters, and response fields** against the [API reference](https://api.reference.langfuse.com) instead of inferring them from code examples.
- **Use the [Langfuse CLI](https://langfuse.com/docs/api-and-data-platform/features/cli)** (`npx @langfuse/cli api <resource> <action>`) to read or write traces, prompts, datasets, and scores from the terminal.

Found an error in these docs? Please open an issue at <https://github.com/langfuse/langfuse-docs/issues>.