OpenAI launches advanced memory system for ChatGPT with ‘Dreaming V3’ upgrade

ChatGPT now remembers more about you, and it does so in a way that’s fundamentally different from how it worked before. OpenAI rolled out its upgraded memory architecture on June 4, branded as “Dreaming V3,” giving the chatbot a significantly improved ability to retain context, maintain conversational continuity, and keep its references to your past interactions feeling relevant rather than stale.

The initial rollout targets ChatGPT Plus and Pro subscribers in the US, with plans to expand to additional subscription tiers and global regions in the near future. Free-tier access is also reportedly in progress.

From sticky notes to an actual brain

The journey started back in February 2024 with a basic “saved memories” feature, where users could tell it to remember specific things, and it would dutifully store them for later.

In April 2025, the model gained the ability to reference all previous chat history, not just the memories users explicitly asked it to save.

The Dreaming V3 system synthesizes context more effectively across those conversations. It doesn’t just recall that you mentioned a project three weeks ago — it understands where that project fits into the broader arc of your ongoing work and adjusts accordingly.

The optimizations focus on three core dimensions: freshness, continuity, and relevance. Freshness means the system prioritizes recent and up-to-date information over outdated context. Continuity ensures threads of conversation feel connected even when separated by days or weeks. Relevance filters out noise, surfacing only the context that actually matters for the current exchange.

User controls and the privacy calculus

OpenAI is giving users the ability to view, edit, or delete their stored memories. For users who want to interact without building up a memory profile, temporary chats offer a full opt-out. Nothing from those sessions gets stored or referenced later.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.