OpenAI blog: Codex brought three open-weight models to high performance in two months

eliebakouch · x · 2026-08-25

OpenAI's blog revealed that using Codex with GPT-Astra, the team brought three open-weight models — not part of Jalapeño's original production plan — to high performance within two months. The poster quotes it and notes the blog also says OpenAI doesn't use MLA for attention, yet the MLA kernel was implemented by Codex without human supervision. He questions whether two months should really count as fast.

Related event: OpenAI Says Codex Tuned Three Open Models in Two Months, Sparking Skepticism(2 posts)→

Original post →

More from coding & agent

coding & agent channel →