Summary & Key Takeaways ​

• llm 0.33 upgrades to the OpenAI Python library 3.x, switching HTTP client dependencies. - The release introduces --key support for llm embed and llm embed-multi commands. - Embedding models now use a consistent pattern for keys, mirroring regular LLM models. - The llm prompt -t/--template option can now be repeated to combine multiple templates. - Reasoning-capable Responses API models now support a reasoning_summary option.

Our Commentary ​

Simon Willison's llm tool continues to be a quiet workhorse for local AI experimentation. This 0.33 release, while not flashy, brings some solid quality-of-life improvements. The ability to combine templates is particularly neat; it feels like a step towards more modular and reusable AI workflows. We appreciate the continuous refinement of these essential developer tools.