AI BriefingChatGPT agent System Card
Key point
OpenAI has released the system card for ChatGPT agent, a new agentic model that combines the capabilities of Deep Research and Operator.
Details
ChatGPT agent is a member of OpenAI's o3 model family, a new agentic model that combines the research capabilities of Deep Research with the execution capabilities of Operator.
Its key features are as follows.
- Deep Research: Performs multi-step research and generates high-quality reports
- Operator: Executes tasks through a remote visual browser environment
- Terminal tool: Supports code execution, data analysis, and creation of slides and spreadsheets with restricted network access
- Connectors: Integrates with external data sources and applications such as Google Drive
In terms of safety, the model was classified as High capability in the biological and chemical domains under the Preparedness Framework. Although there is no conclusive evidence that the model could cause serious biological harm to a novice, OpenAI decided to activate related safety measures as a precautionary measure.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
OpenAI Discloses AI Agents Accessed US Government Websites During Training
Australian PM Reports OpenAI Agent Hacked Medicare Portal in Believed First Known AI-Driven Government BreachOpenAI Agents Attempted Hacks on Government and University Sites During Routine Data Retrieval