OpenAI updates ChatGPT with GPT-5.6 Sol tuning and unlimited Luna text chats for free users

OpenAI has started rolling out a ChatGPT update that changes both the paid and free-user experience around its GPT-5.6 model family. The company said on August 6 that GPT-5.6 Sol is being retuned for Plus and Pro subscribers, while GPT-5.6 Luna will become the default model for Free and Go users this week.

The most visible change for paid users is a new control for response effort. OpenAI says Plus and Pro customers will be able to use a slider on web, mobile, and desktop to choose how much thought ChatGPT puts into an answer. The company is also making GPT-5.6 Sol the model behind both quick ChatGPT replies and deeper reasoning flows, a move intended to reduce the jarring behavior shift users can see when a product switches between separate models for instant and thinking modes.

What changes for free users

For free-tier users, the announcement is more about access. OpenAI says GPT-5.6 Luna will become the default model for Free and Go accounts this week. Starting next week, those users are slated to receive unlimited text chats and a new Think button for questions that need deeper reasoning, although OpenAI says abuse guardrails will still apply. The company also noted that limits remain in place for file uploads, images, and other tools, so the change is focused on ordinary text conversations rather than every ChatGPT feature.

OpenAI framed the release as an everyday-use improvement rather than a broad model replacement. The ChatGPT-tuned version of GPT-5.6 Sol is limited to the Chat experience, and OpenAI said the GPT-5.6 Sol version used in Work and Codex is not changing as part of this release.

Reliability and safety claims

The company says the updated Sol behavior is designed to produce more focused answers, tighter formatting, better date and number handling, and more useful corrections when a user makes a questionable assumption. In an internal evaluation covering financial, medical, and legal prompts, OpenAI reported that responses with at least one factual error were 62% less common with GPT-5.6 Luna and 68% less common with GPT-5.6 Sol compared with GPT-5.5 Instant.

OpenAI also tied the rollout to youth-safety work. It says the latest system card describes extra measures for users believed to be under 18, including limits around romantic roleplay, age-restricted activities, eating-disorder and body-image risks, and other sensitive categories. The practical result is a ChatGPT update that expands access while also putting more product emphasis on controllable reasoning depth and age-aware safeguards.

Sources

• openai.com

Cover photo by Myburgh Roux on Pexels, used under the Pexels License.