OpenAI’s next breakthrough, GPT-6, is now in rapid development. CEO Sam Altman says the model will feature powerful memory capabilities, creating AI assistants that adapt to every user.
OpenAI CEO Sam Altman has confirmed that GPT-6 is entering an accelerated development phase, with a release timeline shorter than the gap between GPT-4 and GPT-5.
“People want memory. People want product features that require us to be able to understand them,” Altman told reporters last week in San Francisco. He called enhanced memory his favorite feature of the year, noting that it will allow ChatGPT to remember preferences, routines, and quirks, making the AI experience more personal.
Altman also revealed that OpenAI is working with psychologists to study how users emotionally engage with AI over time, though the company hasn’t yet released findings.
The new model will focus on personalized AI assistants, powered by a memory function that records user preferences, habits, and quirks to deliver tailored interactions. Altman said GPT-6 aims to move beyond simple Q&A toward deeply adaptive services.
OpenAI is also working with psychology experts to study how AI affects users’ emotions and mental health. At the same time, Altman acknowledged privacy concerns, noting that current temporary memory data still lacks full encryption.
Future versions will comply with the latest U.S. AI executive orders, ensuring neutrality. Altman also hinted at long-term ambitions such as brain-computer interfaces and cross-disciplinary breakthroughs in robotics, energy, and material science.
Despite lukewarm reception to GPT-5, OpenAI quietly refined it with a softer interaction style. GPT-6, however, is being positioned as a game-changer in AI, aiming to become a true personalized digital partner for everyday life.