Skip to main content
These are the models in Ragrails’ built-in LLM catalog. For catalog models, rag.llm(model=...) can infer the provider from the model name. For a custom or newly released provider model, pass both provider= and model=. Pricing is USD per 1M tokens and reflects the package catalog, not a live provider lookup.
Provider model names, prices, and availability change. Verify this catalog against provider docs before publishing release material or quoting prices externally.
Models marked Tools = ✗ don’t support tool calling, so they can’t be used with the interactive CLI’s agentic tools. This includes all Google Gemini models and the OpenAI o-series reasoning models.

​
OpenAI

Built in. Set OPENAI_API_KEY.
The default LLM model in the implementation is gpt-4o. Examples may choose gpt-4o-mini for lower-cost quickstarts.

​
Anthropic

Built in. Set ANTHROPIC_API_KEY.

​
Google Gemini

Built in. Set GEMINI_API_KEY.

​
Embedding models

Set VOYAGE_API_KEY. Install with pip install "ragrails[voyage]".

​
Reranker models

Embedding and reranker providers can be extended with register_embedder() and register_reranker(). Unknown LLM model names can be used when a supported provider is supplied explicitly.