InjazAI

LLM 0.32rc2 changes the default model to GPT-5.6 Luna and adds a new command for interacting with arbitrary OpenAI-compatible endpoints, giving MENA developers a flexible tool for local model testing.

1 min read

llm 0.32rc2 Released: GPT-5.6 Luna Default Model and New CLI Command

Release Summary

FAQ

What is llm 0.32rc2?

It's a release candidate of the LLM CLI tool, featuring GPT-5.6 Luna as default model and a new command for OpenAI-compatible endpoints.

How does GPT-5.6 Luna compare to GPT-4o mini?

Luna is newer and better performing, but slightly more expensive: $0.20 per million input tokens vs $0.15 for 4o mini.

Can MENA developers use this release with local models?

Yes, the llm openai endpoint command allows connecting to local models like LM Studio, enabling testing without cloud costs.

Source: Simon Willison (LLM & tools)

AI-assisted content, human-reviewed.