Free ChatGPT brings unlimited chat

Users of the free version of ChatGPT and the Go package will soon see a significant change. OpenAI has officially announced the removal of text communication restrictions, with the new default model becoming GPT-5.6 Luna, replacing the previous GPT-5.5 Instant. Free unlimited conversations and a new think button will be available the week of August 10th, as OpenAI gradually rolls out the update.

For paid users of the Plus and Pro packages, some of the new features went live on August 6. They received an updated version of the GPT-5.6 Sol model in the interface and a special slider with which they can specify exactly how much time and effort the algorithm should devote to each answer. The lower setting is intended for quick everyday questions, while the higher one allows for more detailed planning, research, or writing code. With a single model for quick and in-depth answers, OpenAI has also eliminated the noticeable differences in the tone of communication that previously appeared. The updated Sol model is currently only available in classic chat, while the ChatGPT Work and Codex platforms remain unchanged.

A similar think button will be added to the free version and Go package the week of August 10. This allows Luna to take more time to think about more complex questions. However, it is worth being a little careful about the term “unlimited”. Free use without restrictions applies only to plain text, while file uploads, image creation and other tools continue to be subject to existing restrictions. OpenAI also points out the security safeguards against abuse, which means that extremely intensive users may still encounter occasional system lockouts.

According to the manufacturer, the new model produces significantly fewer errors. In internal testing of financial, medical, and legal questions, the GPT-5.6 Luna model made about 62 percent fewer errors compared to the GPT-5.5 Instant model, while the Sol model reduced the number of errors by about 68 percent. Independent verification of these numbers is not yet available.

In addition to technical upgrades, OpenAI has also expanded safety restrictions for youth under the age of 18. The model is geared to avoid romantic role-playing, dangerous challenges, and content with dating or inappropriate themes.