LLMs

LLMs, found in the "LLM" section of the dashboard, represent LLMs that you can choose when using various Gabber products. These entities allow you to select and configure different language models to suit your specific needs, whether you are using them for chat completions, real-time sessions, or other applications within the Gabber ecosystem.

Using with /chat/completions

A Gabber LLM Id can be used in place of the model field when using Gabber's OpenAI-compatible chat/completions endpoint.

Using in a Realtime Session

LLM choice is part of the Realtime Session Configuration.

See Generating Connection Details.

Bring your own LLM

If you need to use an LLM that is not supported by Gabber you can bring your own provided it is OpenAI-compatible.

When you bring your own LLM, you can still use it with Gabber features such as the tool calling system and memory contexts.

You will not get metered for tokens consumed and produced by your LLM. You may be metered for tokens consumed within Gabber. For example, tokens used within Gabber's advanced tool calling system.

Bring-your-own-LLM is currently a beta feature, please reach out to the Gabber team to get onboarded.