{
  "id": 12508128,
  "title": "llm-openai-decisions 0.1a0",
  "url": "https://urgent.news/2026/10/06/llm-openai-decisions-0-1a0",
  "topic": "ai",
  "section": "AI",
  "published": "2026-10-06T23:04:13.000Z",
  "source": {
    "name": "Simon Willison",
    "slug": "simon-willison",
    "url": "https://simonwillison.net/2026/Oct/6/llm-openai-decisions/"
  },
  "original_language": "en",
  "account": "OpenAI has unveiled a new Decisions API, known as Jev-style Decisions API, following an announcement at last week's DevDay conference. The company's existing llm-typesafe plugin allows users to interact with the Jev model, and the new plugin, termed llm-openai-decisions, builds upon this foundation. Unlike Jev, the GPT-6 Astra-powered plugin can process image inputs, in addition to textual data.\n\nBoth the Jev and OpenAI Decisions models charge users per million input tokens, with OpenAI's API costing 10 cents per million, while Jev's charges 4.2 cents per million. The API's structure mirrors that of Jev, at least in concept. Both support three types of questions: yes/no, choices, or scores.\n\nAn example of using the new API with an image attachment is provided: one would consult the included README for comprehensive instructions on how to formulate questions of each type. This information comes from a report by Simon Willison, dated October 6th, 2026.",
  "summary": "Release: llm-openai-decisions 0.1a0 OpenAI released their new Jev-style Decisions API , as previously announced at last week's DevDay. Since I already have an llm-typesafe plugin for talking to Jev, I had GPT-6 Astra read the new OpenAI API documentation and build an llm-openai-decisions plugin inspired by llm-typesafe . Unlike Jev, the new gpt-6-luna decision model supports image input in…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}