Universal LLM Model Support (Custom LLM models)

Unique can connect to almost any Large Language Model (LLM), whether it is an open-source model or a proprietary one. Depending on the LLM and the deployment model of Unique, using a non-standard model (e.g., Azure OpenAI) may have certain implications:

Proprietary LLM API Providers

We can connect your Unique deployment to any AI platform of your choice. Examples include OpenAI, Anthropic, and Google Cloud.

You can use any model supported by these platforms. If the account on the platform is managed by Unique, you will benefit from our negotiated rate limits and pricing.

We also optionally support a bring-your-own-API-key model, allowing you to sign up and manage the relationship with the third-party provider yourself. This option is only available for single-tenant, customer-managed tenant, and on-premise deployments.

Open-Source LLM Inference Providers

We also support a number of inference platforms for open-source models. You can use any model they serve via Unique. As with proprietary providers, we offer both Unique-managed accounts and bring-your-own-API-key options.

How to Set Up Spaces Within Unique for Custom LLM Models

Each space in Unique can be configured with multiple AI assistants. Each assistant has its own configuration, which you can adjust directly in the space’s configuration screen:

Open

1. Locate the languageModel parameter within the AI assistant configuration.

2. Define the custom model using the format: litellm:<model-name>

Open

Please contact our support team or refer to your contract to confirm which models are activated for your deployment and the correct name to use in the languageModel parameter.