OpenAI has quietly extended ChatGPT's memory function to users on free plans, marking a broader rollout of a feature once limited to subscribers. The update, which enhances how the chatbot recalls previous conversations, began rolling out this week.
For the first time, people using ChatGPT without a subscription will start noticing more personalized responses as the system starts pulling context from earlier interactions. That means if a user previously asked the AI to avoid certain punctuation or preferred a specific tone, it may now carry that preference into future chats.
The memory system isn't new, but the expanded access is. Originally exclusive to paying users, the upgrade lets free-tier users experience a more conversational and consistent AI—though not without caveats.
According to OpenAI’s updated documentation, free users are getting what could be called a "light" version. This means the AI can retain short-term context across a few recent sessions, but it won’t dig deep into long-past interactions the way it does for paying customers. In contrast, subscribers to the $20/month Plus plan or the $200/year Pro tier receive more robust memory functionality, including long-term recall across a broader history.
For those concerned about privacy or simply preferring one-off interactions, memory can be turned off entirely. Users can head into the settings menu, disable the feature altogether, or selectively erase stored memories.
Alternatively, those who want a clean slate every time can switch to “Temporary Chat” mode—a session that forgets everything once the conversation ends, similar to using a browser in incognito.
This gradual rollout reflects OpenAI's continued push to refine personalization while balancing transparency and user control. As memory improves, so does the potential for context-aware responses—but so does the need for thoughtful use.
Read next: Meta and Yandex Secretly Collected Android Users’ Private Browsing Data Without Permission
For the first time, people using ChatGPT without a subscription will start noticing more personalized responses as the system starts pulling context from earlier interactions. That means if a user previously asked the AI to avoid certain punctuation or preferred a specific tone, it may now carry that preference into future chats.
The memory system isn't new, but the expanded access is. Originally exclusive to paying users, the upgrade lets free-tier users experience a more conversational and consistent AI—though not without caveats.
According to OpenAI’s updated documentation, free users are getting what could be called a "light" version. This means the AI can retain short-term context across a few recent sessions, but it won’t dig deep into long-past interactions the way it does for paying customers. In contrast, subscribers to the $20/month Plus plan or the $200/year Pro tier receive more robust memory functionality, including long-term recall across a broader history.
For those concerned about privacy or simply preferring one-off interactions, memory can be turned off entirely. Users can head into the settings menu, disable the feature altogether, or selectively erase stored memories.
Alternatively, those who want a clean slate every time can switch to “Temporary Chat” mode—a session that forgets everything once the conversation ends, similar to using a browser in incognito.
This gradual rollout reflects OpenAI's continued push to refine personalization while balancing transparency and user control. As memory improves, so does the potential for context-aware responses—but so does the need for thoughtful use.
Read next: Meta and Yandex Secretly Collected Android Users’ Private Browsing Data Without Permission
