OpenAI is reportedly preparing to release a new, highly anticipated version of its chatbot, ChatGPT-5, potentially as early as August, though delays until the end of the month are possible. This information was reported by The Verge editor Tom Warren, citing sources close to the company.

In parallel with the primary version, OpenAI is also developing more compact variants: ChatGPT-5-mini and ChatGPT-5-nano. The first two models (main and mini) will be available via subscription and through the API (for developers), while the nano version will be accessible exclusively via the API.

OpenAI CEO Sam Altman has already shared his personal experience using ChatGPT-5, recounting how the model assisted him with a complex question, leading him to feel "useless in the face of AI."

According to Warren, OpenAI aims to declare one of its future models as Artificial General Intelligence (AGI) as soon as possible. This is significant because the current contract with Microsoft permits Microsoft to utilize OpenAI models only until the AGI level is achieved. Rapid attainment of this status could help OpenAI free itself from certain obligations to Microsoft.

The main version of ChatGPT-5 will be a "unified" model, capable of independently selecting its response style: from a quick response similar to ChatGPT-4o to a more in-depth and detailed one akin to ChatGPT-3.5-o (o3). Sam Altman previously promised that ChatGPT-5 would be available to all users, including a free version, but subscribers would gain access to more extensive and higher-quality reasoning from the model.

ORIENT