Add a custom model
Use any OpenAI-compatible endpoint or self-hosted model.
Bring your own model: a self-hosted LLM, a fine-tune, or any provider with an OpenAI-compatible API. Once added, it appears in the model selector next to the built-in models.
Add a custom model
- Open Models in the left icon bar and click + Add Model.
- Fill in:
| Field | What to enter |
|---|---|
| Name | How it appears in the model selector |
| Model ID | The exact model name your API expects, e.g. llama-3.1-70b-instruct |
| Base URL | Your API's base URL, e.g. https://my-llm.example.com/v1 |
| API Key | The key for that endpoint |
| Max Context Length | The model's context window in tokens, e.g. 4096 |
| Description | Optional notes for you and your team |
- Save, then pick it from the model selector.
Works with
Any server that speaks the OpenAI API, for example local runners and hosted inference services that offer an OpenAI-compatible endpoint.
Good to know
- Custom models are available on paid plans.
- Tools, vision and thinking mode depend on what your model supports.
- Your custom models are listed under Your Models; the Pre-Built Models table below lists every built-in model by provider, with vision, tool support, context length and pricing per 1M tokens.
Was this article helpful?