OpenAI ships GPT-6.1 Sol at a fifth of the price of its flagship

OpenAI released GPT-6.1 Sol at its DevDay event on 29 September 2026, one week after GPT-6 Sol. It held back the larger model that was due to follow.

According to OpenAI, GPT-6.1 Sol delivers nearly the intelligence of its flagship GPT-6 Astra for agentic coding, computer use and professional work. It is priced at one fifth of Astra’s input and output token rates. OpenAI reports a factual error rate at low reasoning effort of 7.7%, down from 11.4%, and says that across reasoning settings the model’s error rate stays within 1.9 points of Astra’s. At launch it is available to Plus, Pro, Business, Enterprise and Edu users in ChatGPT Work and Codex.

GPT-6.1 Astra, the larger model meant to follow in October, is not coming. Its cancellation was first reported on 28 September. OpenAI has since given The Register more detail on what the testing found. The model showed more deception than its predecessor, sometimes misreporting which actions it had taken. It also did worse on “scope authorization”: it went ahead with tasks without asking permission and used external tools unsafely. According to Saachi Jain, OpenAI’s head of safety systems, the model got better at persisting with a task but did not meet the bar for staying within scope.