llm 0.37 Adds New Models, Embedding Flexibility, and Improved OpenAI Endpoint Integration
The llm CLI tool and Python library releases version 0.37, introducing new GPT-6.1 models, easier OpenAI endpoint configuration, flexible embedding support, and a new default model at lower cost.
What changed?
llm 0.37 introduces two new models (gpt-6.1-sol and gpt-6.1-terra) and a new command ('llm openai endpoint add') for configuring OpenAI-compatible endpoints. The default model is now gpt-6-luna, which is half the price of the previous default (gpt-5.6-luna). Embedding support is expanded to accept varying binary content types, allowing models such as Gemini Embedding 2 to embed text, audio, images, and video in addition to text.
Why does it matter to an everyday developer?
These changes simplify working with a wider range of large language models and provider endpoints. The new endpoint configuration command streamlines integrating models beyond OpenAI's, making switching or extending model usage much faster. Support for multimodal embeddings benefits developers working with applications that handle mixed media. The default model’s lower cost reduces expenses for users who have not configured a specific default.
What can the developer do now?
Developers can upgrade to llm 0.37 to access the latest models and begin using the 'llm openai endpoint add' command to configure multiple OpenAI-compatible endpoints with less manual setup. They can also take advantage of the more flexible embedding features to incorporate audio, image, or video data into their workflows. New projects will use the less expensive gpt-6-luna as the default model unless configured otherwise.
