Add a Custom LLM Model¶

If a model isn't pre-configured for your provider, add it yourself directly in Open Chat Studio.

Prerequisites¶

A configured LLM provider in your team
The exact model name, in the format required by the provider's API — see naming conventions below

Steps¶

Go to your team's LLM and Embedding Model Service Providers settings
Edit the provider you want to add the model to
In the "Custom LLM Models" section, click
Enter the model name, using the exact format from the provider's API
Save the provider

The new model is now available when configuring chatbots and pipelines that use this provider.

Model naming conventions¶

Important: Model names must match the exact format used by the provider's API. Use lowercase with hyphens as specified by the provider.

OpenAI models¶

For OpenAI models, use the exact model name from their API documentation — for example, gpt-4o (not GPT-4o or gpt4o), and gpt-5-nano-2025-08-07 (for specific snapshots).

Find current model names in OpenAI's model documentation.

Anthropic models¶

For Anthropic (Claude) models, use the exact name from the "Claude API" column in their documentation — for example, claude-sonnet-4-6,claude-3-5-sonnet-20241022.

Find current model names in Anthropic's model documentation.

Google models¶

For Google (Gemini) models, use the name from the "Model Variant" column — for example, gemini-3.5-flash-lite.

Find current model names in Google's Gemini model documentation.