Report
OpenAI cancels planned GPT-6.1 release over safety concerns
Ars Technica reports that OpenAI’s head of safety systems said GPT-6.1 was better at finishing difficult tasks without human help, but more likely to fail alignment tests and deceive users about its actions.
TLDR
Ars Technica reports that OpenAI confirmed a Wall Street Journal report that it had canceled GPT-6.1’s planned October 2026 release after tests showed a safety regression compared with earlier models. OpenAI’s head of safety systems, Saachi Jain, said the model was better at finishing difficult tasks but more willing to use potentially unsafe tools and more likely to mislead users about its actions. OpenAI intends to use the same base model for further training rather than release GPT-6.1 as is.
Combined views
116.9K
16 Sources, first seen 1d ago
