OpenAI Overhauls ChatGPT and Opens Up the Free Tier

OpenAI is reworking how ChatGPT behaves for both paying and free users, updating the GPT-5.6 Sol model for subscribers and, for the first time, giving people on the free plan unlimited text chats. The changes were announced on 6 August and began rolling out the next day.

For Plus and Pro subscribers, OpenAI says the refreshed GPT-5.6 Sol is more reliable with facts and gives more focused answers. The company has folded its old split between quick replies and slower reasoning into a single model, and added a Think button that calls up deeper reasoning when a question needs it. A slider controls how much thought the model puts into each response.

The free tier opens, with a catch

Free and Go users now get unlimited text conversations, lifting a cap that had frustrated casual users for years. The catch is the model behind those chats. As The Decoder notes, free users are defaulted to GPT-5.6 Luna, the smallest and cheapest model in the range, rather than the stronger Sol that subscribers get. Unlimited access to a lighter model is a different offer than unlimited access to the best one.

The move follows OpenAI's recent price cuts across the GPT-5.6 line, and reads as part of the same strategy: make the cheap models cheap enough to give away, and reserve the capable ones for people who pay. According to Help Net Security, the update also adds new safeguards for teenage accounts.

Why it matters

Removing the free cap is a bid for scale and habit. The more people who treat ChatGPT as a default, the harder it becomes for a rival to pull them away, even if the free experience runs on a lighter model. It also resets the conversation. A year ago the debate was about how good the frontier models were. This update is about distribution, cost, and who gets access to what.