
OpenAI’s Decisions API returns typed answers; the company says it is 10x faster than Responses
The public beta uses only gpt-6-luna and bills input tokens. It is for predicates, choices and scores, not generated objects.
OpenAI news from AiLookout: the latest ChatGPT features, GPT model releases, Codex updates, funding, data centers and policy, each linked to its source.

The public beta uses only gpt-6-luna and bills input tokens. It is for predicates, choices and scores, not generated objects.

Researchers poisoned a Qwen2.5 fine-tune for under $50. The fire rates are their own unreplicated results, not a general exploit recipe.

The 30 September post attributes a core cluster to individuals linked to Kimi’s developer. It does not show that extracted reasoning trained a Kimi model.

The 9 October Sierra post publishes a draft of Personal Agent Protocol, called Poppy, and lists 35 more design partners including OpenAI. The v0.1 text is at personalagentprotocol.org under Apache 2.0. A reference implementation is still promised.

Paid ChatGPT users can upload common audio files up to 512 MB, but OpenAI warns that transcripts and speaker labels can be wrong.

The expanded partnership connects OpenAI models with Atlassian’s permission-aware Teamwork Graph, while deeper Jira agent workflows remain future work.

OpenAI posted 722 manuscripts on 6 October. On 7 October its history.md withdrew three after a sign error and said 300 of 719 top-line results were formalized. We have not checked the Lean kernels.

An 11-task research evaluation measured whether an AI agent could configure agreements, approvals, clauses, and procurement workflows inside Ironclad.

OpenAI plans a US test later this month and expands campaign measurement.

API customers can opt in now. OpenAI warns that detection has limits.

Albertsons’ expanded OpenAI partnership connects conversational shopping with a familiar retail boundary: the final cart and purchase.

OpenAI’s new model puts capability and cost in the same conversation. Here’s what the announcement means for builders.

Always-on agents bring a new set of questions about access, responsibility, and how we measure useful work.

A look at the developer announcements, from cloud coding environments to tools for ongoing responsibilities.

ChatGPT vs Claude vs Gemini depends on the task, the sources, and the account you have. Compare the finished work, not the model name.