Across 20 behavior categories and three GPT-5-series Thinking deployments, simulated and observed rates were strongly correlated. The method outperformed challenging-prompt and previous-deployment baselines at predicting whether rates would rise or fall—and by how much.
@openai
-
Analysis of ChatGPT conversations from opted-in users with anonymization and aggregate results
By
–
For this research, we analyzed only ChatGPT conversations from users who allow their data to be used to improve models. Before analysis, we removed account-linked identifiers and identifiable information, and we report only aggregate findings.
-
OpenAI analyzes opt-in ChatGPT conversations after anonymization
By
–
For this research, we analyzed only ChatGPT conversations from users who allow their data to be used to improve models. Before analysis, we removed account-linked identifiers and identifiable information, and we report only aggregate findings.
-
Simulating deployment to predict model behavior before release
By
–
We’re sharing new research on a method for anticipating how models may behave in real-world use before release: simulating deployment with recent, de-identified user requests and studying candidate model responses.
-
Traditional evaluations and deployment simulation for AI risk assessment
By
–
Traditional evaluations and red-teaming remain essential, especially for rare or severe risks. Deployment Simulation complements them by helping us estimate how often undesired behaviors may occur in realistic use and surface new behaviors before release.
-
Model finds counterexample to 80-year-old Erdős conjecture
By
–
What happened when one of our models found a counterexample to an 80-year-old Erdős conjecture?
— OpenAI (@OpenAI) 4 juin 2026
Researchers @alexwei_, @HongxunWu, and @wjmzbmr1 shared the story on the OpenAI Podcast with @AndrewMayne and explained how mathematicians and models can work together to make new… pic.twitter.com/bQQ6Bvr8QhWhat happened when one of our models found a counterexample to an 80-year-old Erdős conjecture? Researchers @alexwei_
, @HongxunWu
, and @wjmzbmr1 shared the story on the OpenAI Podcast with @AndrewMayne and explained how mathematicians and models can work together to make new -
New memory system enables review and control of ChatGPT’s context
By
–
With the new memory system, you can review and steer what ChatGPT remembers through a memory summary, with more visibility and control over how context is used. pic.twitter.com/kXMAds0g3q
— OpenAI (@OpenAI) 4 juin 2026With the new memory system, you can review and steer what ChatGPT remembers through a memory summary, with more visibility and control over how context is used.
-
ChatGPT memory adapts to trip planning context and preferences
By
–
We’re building ChatGPT to remember what matters, follow your preferences and constraints, and adapt as things change. If you tell ChatGPT you’re planning a trip in July, memory should understand when the trip is upcoming, happening, and already over. That helps ChatGPT keep
-
OpenAI adds new capabilities to GPT-Rosalind for drug discovery
By
–
We’re bringing new capabilities to GPT-Rosalind, a model series purpose-built for life sciences research at enterprise scale. It brings GPT-5.5’s agentic coding and tool use together with stronger intelligence for drug discovery, analysis, design, and experimental workflows.
-
Codex Plugin for Public Equity Investing from Questions
By
–
From question to model.
— OpenAI (@OpenAI) 2 juin 2026
The public equity investing plugin for Codex. pic.twitter.com/RCigvKfNfcFrom question to model. The public equity investing plugin for Codex.