OpenAI has expanded its developer platform with two “deep research” models—o3-deep-research-2025-06-26 and o4-mini-deep-research—designed for complex analytical tasks that require access to private or proprietary data. The company simultaneously enabled native web-search functionality in the existing o3, o3-pro and o4-mini models, allowing them to retrieve real-time information as part of their reasoning processes. Web-search calls now cost US$10 for every 1,000 queries, a price cut intended to spur broader adoption. The new deep-research endpoints are priced at US$10 per one million input tokens and US$40 per one million output tokens, making a single intensive query potentially run to tens of dollars, according to early user tallies. The release also introduces webhooks for API event notifications, giving developers a way to trigger downstream actions whenever a model completes a task. OpenAI said further details will be outlined at its DevDay 2025 conference as it courts enterprise and independent developers looking to build more sophisticated AI workflows.