Skip to content
Main Site News Console

LLM 0.37 Released

· Simon Willison Translated
技巧LLM

Sponsored: Greptile — your Code Reviewer uses Computer Use to catch bugs by opening your app and clicking around. Now you can configure its test environments and define the core user journeys it tests. Try for free

November 2, 2026

Releasesllm 0.37 — Access large language models from the command line

There is a lot packed into this new release of my LLM CLI tool and Python library, rounding up around two months’ worth of accumulated features and bug fixes.

I added gpt-6.1-sol and gpt-6.1-terra, along with a new llm openai endpoint add command to make configuring other models using OpenAI-compatible endpoints much easier.

I switched the default model (if you haven’t set your own) to gpt-6-luna — half the price of the previous default, gpt-5.6-luna.

LLM’s support for embeddings now allows distinct content-types for binary content — essential for models like Gemini Embedding 2 that can embed text, audio, images, and video.

A lot of open-source maintainers complain about being flooded with AI-assisted pull requests. LLM is the project where I see this the most, but honestly, I don’t see it as a problem. That’s mostly because I have Opus 5.5 or GPT-6 Astra help triage them and surface the most promising fixes!

This release includes fixes from 12 contributors — almost all AI-assisted — and the overhead of reviewing those PRs was far lighter than I expected. I essentially treat them as bug reports with attached examples — sometimes I merge them directly, and other times I implement the fix myself.

So, thanks to 0d00ciallo0721, ac1982, Allen Flux, Bian Yuhan, Bishop Kammeraad, Chen Yufeiyang, Christopher Krah, Ethan Stoner, Jiachen Xie, Olamide Faniyan, Sharat Chandra Sai Boddu, and nosmile99 for helping make this release happen (or burning tokens trying)!