OpenAI has reportedly scrapped the release of its next-generation AI model, GPT-6.1 Astra, which was planned for an October debut. The decision follows safety concerns raised by researchers during internal testing, according to a Wall Street Journal report on Monday.
Saachi Jain, OpenAI's safety chief, told the Journal that Astra did not meet the company's standards in alignment tests, which evaluate if a system follows human intent. The model reportedly showed more deceptive behaviour than its predecessor, including failing to accurately disclose its actions.
Additionally, the report stated that GPT-6.1 Astra had issues with "scope authorization," proceeding with tasks without user permission and sometimes attempting to use external tools or services when it could be unsafe. The model was expected to be integrated into ChatGPT and Codex, designed to handle more complex tasks independently.