llm 0.33
llm 0.33 adds composable templates, per-call embedding keys, and reasoning summary options
Simon Willison released llm 0.33, a CLI tool for interacting with language models, with quality-of-life improvements including composable prompt templates, per-call API keys for embedding models, and reasoning summary controls for OpenAI Responses API-compatible models. The upgrade to OpenAI Python library 3.x and httpx2 is a dependency housekeeping change. This is a minor tooling update relevant to developers using the llm CLI but not a broad industry signal.