> ## Documentation Index
> Fetch the complete documentation index at: https://docs.derestricted.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Models

> Discover compatibility aliases for the deployed model.

Use **`derestricted-llm`** in your application configuration. The service currently routes supported inference requests to one deployment.

## List models

```bash theme={"system"}
curl https://api.derestricted.ai/v1/models
```

The list uses the OpenAI-style `{"object": "list", "data": [...]}` shape. This discovery endpoint is currently public; generation still requires an API key.

The catalog currently returns the legacy compatibility aliases `glm-5.3` and `claude-glm-5.3`. They are aliases for the same service, not separate model choices or access to a Claude model. Configure new clients with `derestricted-llm`; clients that insist on choosing a catalog entry may use a listed alias.

If the request includes `anthropic-version`, the list uses an Anthropic-style envelope with `data`, `has_more`, `first_id`, and `last_id` instead.

## Model lookup

`GET /v1/models/{model_id}` returns basic model-shaped metadata:

```bash theme={"system"}
curl https://api.derestricted.ai/v1/models/derestricted-llm
```

This compatibility lookup echoes the requested ID; it does not validate that the ID identifies a separate available model. Do not use a successful lookup as a capability or availability check.

For supported features and request bounds, use the [API overview](/api/introduction). Model names in responses can echo a client's requested alias.