PANews reported on July 18 that OpenAI released a new ChatGPT agent feature in a live broadcast early this morning. The new feature integrates the capabilities of Operator and deep research tools, and supports autonomous completion of complex tasks such as scheduling, competitive product analysis, and presentation creation in a virtual computer environment. Users can directly instruct the agent to browse websites, analyze data, generate slides and spreadsheets in the conversation. The feature is equipped with a variety of tools such as visual browsers, text browsers, terminals, and API access, and has achieved industry-leading results in multiple academic assessments. Pro users can use it almost unlimited times per month, Plus and Team users can use it 50 times per month, and corporate and educational users will gradually gain access in the coming weeks.