OpenAI Revamps ChatGPT Memory, Cuts Computing to One-Fifth to Fix Obsolescence and Errors
OpenAI, the leading AI company, has announced a major architectural upgrade to ChatGPT's memory function, introducing a new system powered by the revolutionary Dreaming V3 mechanism. The core goal of this upgrade is to eliminate two longstanding issues that have plagued users: outdated memory and insufficient accuracy. It also significantly boosts scalability for serving hundreds of millions of users.

Intelligent Evolution Without Explicit Prompts
Looking back at the evolution of ChatGPT's memory system: when it first launched in 2024, it relied entirely on explicit user prompts like "remember something." The system could only mechanically store a limited amount of background information. With the introduction of the Dreaming mechanism in 2025, the system began to autonomously reference chat history and select relevant content through background processes. However, it wasn't until the release of this V3 version that the mechanism truly matured and became the core architecture of the entire memory system.

In the new architecture, ChatGPT has made significant strides in personalization, automatically updating user profiles over time. The system summarizes important life details and preferences, displaying them on a dedicated page. Users can intuitively view, modify, or add this information, and can also ask about the source and details of a specific memory through conversation.
Reduced Compute Power Empowers Free Users
In addition to user experience improvements, OpenAI has also achieved breakthroughs in operational mechanisms and resource optimization. Official data shows that after recent deep technical improvements, the computing resources required by the new Dreaming V3 mechanism for serving free users have been reduced to just one-fifth of the original amount.
Thanks to this sharp reduction in compute consumption, OpenAI has finally been able to extend this advanced memory feature to free users while maintaining service quality. Currently, the system has been initially rolled out to Plus and Pro paid users in the United States and has simultaneously doubled their memory capacity limit.
Related article
ByteDance’s Seed launches global campus drive, offering virtual shares to win top large model talent
In the competitive landscape of large language models, securing top-tier talent remains the most critical strategic asset.On April 1st, ByteDance announced the launch of its Seed global campus recruitment initiative, part of its large model talent de
Suno to Watermark Songs Amid Legal Battles
Suno, the platform enabling users to generate AI-created music, has unveiled new features to label platform-produced tracks, restrict downloads, and update community standards to curb unauthorized replicas. These updates arrive as Suno confronts mult
Musk Admits Grok Build Leaked User Code, Promises to Erase All Historical Data
Elon Musk directly addressed the privacy controversy surrounding Grok Build, beginning with a simple "True" to confirm the incident's validity. He pledged that all user data previously uploaded to SpaceXAI would be permanently erased, stating, "not a
Related Special Topic Recommendations
Comments (1)
0/500
OpenAI, the leading AI company, has announced a major architectural upgrade to ChatGPT's memory function, introducing a new system powered by the revolutionary Dreaming V3 mechanism. The core goal of this upgrade is to eliminate two longstanding issues that have plagued users: outdated memory and insufficient accuracy. It also significantly boosts scalability for serving hundreds of millions of users.

Intelligent Evolution Without Explicit Prompts
Looking back at the evolution of ChatGPT's memory system: when it first launched in 2024, it relied entirely on explicit user prompts like "remember something." The system could only mechanically store a limited amount of background information. With the introduction of the Dreaming mechanism in 2025, the system began to autonomously reference chat history and select relevant content through background processes. However, it wasn't until the release of this V3 version that the mechanism truly matured and became the core architecture of the entire memory system.

In the new architecture, ChatGPT has made significant strides in personalization, automatically updating user profiles over time. The system summarizes important life details and preferences, displaying them on a dedicated page. Users can intuitively view, modify, or add this information, and can also ask about the source and details of a specific memory through conversation.
Reduced Compute Power Empowers Free Users
In addition to user experience improvements, OpenAI has also achieved breakthroughs in operational mechanisms and resource optimization. Official data shows that after recent deep technical improvements, the computing resources required by the new Dreaming V3 mechanism for serving free users have been reduced to just one-fifth of the original amount.
Thanks to this sharp reduction in compute consumption, OpenAI has finally been able to extend this advanced memory feature to free users while maintaining service quality. Currently, the system has been initially rolled out to Plus and Pro paid users in the United States and has simultaneously doubled their memory capacity limit.
ByteDance’s Seed launches global campus drive, offering virtual shares to win top large model talent
In the competitive landscape of large language models, securing top-tier talent remains the most critical strategic asset.On April 1st, ByteDance announced the launch of its Seed global campus recruitment initiative, part of its large model talent de
Suno to Watermark Songs Amid Legal Battles
Suno, the platform enabling users to generate AI-created music, has unveiled new features to label platform-produced tracks, restrict downloads, and update community standards to curb unauthorized replicas. These updates arrive as Suno confronts mult
Musk Admits Grok Build Leaked User Code, Promises to Erase All Historical Data
Elon Musk directly addressed the privacy controversy surrounding Grok Build, beginning with a simple "True" to confirm the incident's validity. He pledged that all user data previously uploaded to SpaceXAI would be permanently erased, stating, "not a





Home






