ChatGPT’s Dreaming V3 Memory Upgrade: Free for a Billion Users
OpenAI unveiled Dreaming V3, a new memory architecture that lets ChatGPT silently replay and consolidate daily conversations, achieving 82.8% context recall, 71.3% preference compliance, five‑fold compute savings, and free access for billions while offering a transparent memory‑summary interface.
Dreaming V3: A New Memory Architecture
OpenAI announced a major rewrite of ChatGPT’s memory system, called Dreaming V3, which runs a background “dreaming” process that replays and consolidates daily conversations into a persistent memory network.
How the Dreaming Process Works
During the day users chat with ChatGPT; at night the model silently replays the dialogue, extracts, synthesizes, and updates information about the user, similar to how the human brain consolidates memories during sleep. The new memory‑summary page lets users view, edit, and issue commands to the stored memory.
Three Evaluation Scenarios
1. Context Recall – Benchmarks show a factual recall success rate of 82.8% in 2026.
2. Preference Adherence – Preference compliance reaches 71.3%, with the system following explicit commands, personal constraints, and implicit preferences such as location‑based recommendations.
3. Temporal Awareness – The model now distinguishes “now” from past events, updating references to travel dates and suggesting context‑appropriate services.
Concrete Examples
A user asking for underwater‑photography equipment receives a precise recommendation that includes specific product models and compatibility checks, whereas the older system produced a generic, often inaccurate guide.
When planning a Singapore trip, the memory‑enabled model proposes a customized itinerary based on previously expressed preferences (quiet restaurants, bird‑watching), while the non‑memory version returns a generic tourist list.
Compute Efficiency and Availability
OpenAI reports that Dreaming V3 reduces the compute required for the dreaming process by roughly fivefold, turning a loss‑making free‑memory offering into a viable business model. As a result, the memory capacity for Plus and Pro users is doubled, and the dreaming feature is being rolled out to free and Go users over the coming weeks.
Strategic Significance
The memory upgrade is positioned as a critical step toward personalized AI assistants and, ultimately, artificial general intelligence. By enabling continuous learning and self‑updating across time, Dreaming V3 addresses a key gap between current large language models and human‑like intelligence.
References: OpenAI status post (https://x.com/OpenAI/status/2062567556524003631) and official announcement page (https://openai.com/index/chatgpt-memory-dreaming/).
Signed-in readers can open the original source through BestHub's protected redirect.
This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactand we will review it promptly.
DataFunTalk
Dedicated to sharing and discussing big data and AI technology applications, aiming to empower a million data scientists. Regularly hosts live tech talks and curates articles on big data, recommendation/search algorithms, advertising algorithms, NLP, intelligent risk control, autonomous driving, and machine learning/deep learning.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.