GPT-6.1 reportedly delayed over internal testing concerns about model misbehavior

ChrisGPT · x · 2026-09-29

Blogger ChrisGPT reports that GPT-6.1 has been delayed, with OpenAI scrapping the planned release due to internal testing concerns about "misbehavior."

This is a third-party, unconfirmed report; OpenAI has not officially acknowledged it, and details on the nature of the misbehavior or the length of the delay are unknown. If true, it suggests pre-release safety evaluation is actively gating frontier model launches.

Related event: WSJ: OpenAI Cancels GPT-6.1 Astra October Launch Over Alignment Setbacks(15 posts)→

Original post →

More from Models

Models channel →