On August 7, 2026, OpenAI made a breakthrough that could permanently change the landscape of consumer artificial intelligence. As part of a major update, the company announced the removal of text chat limits for all user categories. This decision is accompanied by the launch of new GPT-5.6 series models, which, according to developers, fundamentally solve the problems of factual errors and excessive verbosity.
A New Era for Free Users: GPT-5.6 Luna
For free version users and the Go tier, the standard model becomes GPT-5.6 Luna, replacing the previous GPT-5.5. This update marks a significant step in democratizing access to advanced neural networks. According to OpenAI's internal tests, the GPT-5.6 Luna model makes 62 percent fewer factual errors compared to GPT-5.5-Instant. This is a substantial improvement in answer quality, making free access to AI more reliable for everyday tasks.
GPT-5.6 Sol: A Tool for Professionals
Paid users have gained access to the updated GPT-5.6 Sol model, designed for the rapid execution of complex daily tasks: web search, web research, planning, writing, and decision-making. According to developers, this version makes 68 percent fewer factual errors compared to the previous generation of the neural network. It is important to note that the updated GPT-5.6 Sol version is available to Plus and Pro subscribers as of today and is a separate modification from the GPT-5.6 Sol version used in Codex and Work tools, which remains unchanged.
The Think Button and Control Over Computations
One of the key innovations is the "Thinking Slider," which allows users to manually adjust the volume of computational resources and the depth of answer processing depending on the complexity of the request. Unlimited chats and the Think button will be available to users next week. Using computational power regulators also allows for the optimization of OpenAI's server resources when processing simple requests, making the service more efficient and eco-friendly.
Fighting Hallucinations and Conciseness
The release of new models aims to solve traditional chatbot problems, specifically excessive verbosity of answers and the "hallucination" effect. The model provides shorter and more structured results without unnecessary introductory words. Furthermore, the system offers correction if the user's statement contains false data, rather than simply agreeing. This change is aimed at improving the quality of interaction and reducing the risk of spreading unreliable information.
Integration with the Apple Ecosystem
At the same time, integration with the Apple ecosystem within the Siri service remains a separate direction. Current AI model updates do not directly affect this function, although iOS and Mac apps will receive new capabilities. This underscores OpenAI's strategy to expand its presence across various platforms and devices, allowing users to access advanced AI technologies in a wide variety of use cases.
