The update improves GPT-5.6 Sol for everyday chat with more direct answers, fewer factual mistakes, and a more consistent tone between instant and deeper-reasoning modes. OpenAI also introduced a thought-effort slider for paid users and expanded GPT-5.6 Luna availability for free users.
Even though this release is framed around ChatGPT, it matters to developers who rely on chat for planning, debugging, research, and lightweight coding support. The bigger shift is product packaging: OpenAI is making stronger reasoning controls and higher-utility defaults available to a wider user base without forcing a separate model workflow.
Paid users can open ChatGPT on web, desktop, or mobile and adjust the new slider when moving from quick lookups to planning, writing, or coding tasks. Free users can use GPT-5.6 Luna as the default chat model and press Think when a prompt needs more deliberate reasoning.
Read Original Post →