llm 0.33

Simon Willison has upgraded the OpenAI Python library to version 3.x and switched the HTTP client dependency from httpx to httpx2. A comprehensive fix for a shipped 0.32.1 fix was released, which allows llm embed and llm embed-multi to accept --key. The Python EmbeddingModel.embed(), EmbeddingModel.embed_multi(), Collection.embed() and Collection.embed_multi() methods now accept key= too, passing the resolved per-call key to embedding plugins without changing shared model state.

Existing plugins that read self.key continue to work through a compatibility fallback. The embedding models now use the same pattern for keys that regular LLM models do. The llm prompt -t/--template can now be repeated to combine templates in order, unlocking a pattern where model configuration and options from one template can be used with a prompt from another.

OpenAI's Reasoning-capable Responses API models now support a reasoning_summary option with auto, concise, and detailed values, which can be used with llm openai endpoint --responses.

Brief written by urgent.news from Simon Willison's own syndicated text. Machine-written — may contain errors; check the original before relying on it.