OpenAI is updating ChatGPT to improve factual reliability for Plus and Pro users while expanding access for free users. The update introduces a new slider allowing users to control the reasoning effort and changes the default model for free accounts.

  • GPT-5.6 Sol in Chat now provides more focused answers and fewer factual errors, with internal evaluations showing 62% fewer errors for Luna and 68% fewer for Sol compared to GPT-5.5 Instant.
  • Plus and Pro users gain a new slider to adjust how much thought the model puts into responses, unifying Instant and Thinking experiences.
  • GPT-5.6 Luna becomes the default model for Free and Go users this week, accompanied by unlimited text chats and a new Think button for deeper reasoning starting next week.

These changes aim to make the latest models more widely available and improve the usefulness of answers across all user tiers.