OpenAI has introduced several new features in its API, including two deep research models named o3-deep-research and o4-mini-deep-research, designed to enhance complex research capabilities by breaking down questions, performing searches, running code, and citing sources. These models are priced at $10 to $40 per million input/output tokens. Additionally, OpenAI has integrated web search functionality into its o3, o3-pro, and o4-mini models, allowing real-time web searches within the model's reasoning process at a reduced price of $10 per 1,000 tool calls. The update also includes webhooks for API event notifications. These advancements aim to enable developers to build more sophisticated and scalable AI workflows, combining deep research with writing, coding, and media generation in unified chat environments. The features have been adopted by platforms such as Poe, which now supports switching between deep research and other AI models seamlessly.
OpenAI’s Deep Research API is a game changer: ask a question, it breaks it down, searches, runs code, cites sources. No fluff, just evidence-backed insight. #AIResearch https://t.co/F0WScR4uWI https://t.co/7lW0fH39xx
OpenAI Deepens AI's Research Abilities OpenAI has rolled out new "deep research" models in its API, alongside webhooks and enhanced web search, aiming to empower developers to build more sophisticated and scalable AI workflows. What new frontiers of knowledge will be unlocked
There are 100’s of educational AI tools Missing piece is instant DeepResearch for learning So far only a handful are getting there