Willison releases LLM plugin for OpenAI's Decisions API
TL;DR
- Simon Willison's llm-openai-decisions 0.1a0 wraps OpenAI's new Decisions API, announced at last week's DevDay, as a plugin for his `llm` CLI.
- The gpt-6-luna model handles yes/no, choices, and scoring questions over text or image input, where Jev's equivalent remains text-only.
- OpenAI charges 10 cents per million input tokens; Jev's decision model runs at 4.2 cents, and both bill on input only.
Simon Willison has released llm-openai-decisions 0.1a0, a plugin that wires OpenAI's new Decisions API into his `llm` command-line tool. OpenAI announced the API at last week's DevDay, following Jev's own launch of the decision-model shape in September.
Willison wrote the plugin by having GPT-6 Astra "read the new OpenAI API documentation and build an `llm-openai-decisions` plugin inspired by `llm-typesafe`," his earlier wrapper for Jev. The new `gpt-6-luna` model accepts both text and image input, where Jev remains text-only, and both APIs support three question types for yes/no, choices, or scores.
The pricing is the sharper split. OpenAI charges 10 cents per million input tokens; Jev charges 4.2 cents. Both bill on input only.
The release is tagged 0.1a0, an alpha. Documentation lives in the project's GitHub README.
Shared on Bluesky by 2 AI experts
-
Simon Willison @simonwillison.net
I released an LLM plugin for sending prompts to OpenAI's new Jev-clone decision model that's based on GPT-6 Luna simonwillison.net/2026/Oct/6/l...
View on Bluesky →
Originally reported by simonwillison.net
Read the original article →Original headline: Release: llm-openai-decisions 0.1a0