👉 New OpenAI API Features: Deep Research Models, Webhooks, and Web Search with Reasoning We’ve introduced several new capabilities to the OpenAI API to help you build smarter, more scalable AI workflows: 🔍 Deep Research Models in API You can now access the o3-deep-research and o4-mini-deep-research models directly through the API—the same ones available in ChatGPT. These models support both internal knowledge and external context via web search and MCP server integration. o3-deep-research: $10 per 1M input tokens, $40 per 1M output tokens o4-mini-deep-research: $2 per 1M input tokens, $8 per 1M output tokens Use them to programmatically launch in-depth research from your own tools or workflows. 📡 Webhooks for Async Work You can now subscribe to webhooks to get notified when tasks complete—ideal for long-running jobs like deep research or o3-pro tasks. This removes the need for polling and improves reliability when used with background mode. 🌐 Web Search in Reasoning Models Web search is now available in o3, o3-pro, and o4-mini models, allowing them to dynamically fetch and incorporate real-time context into their reasoning. Pricing: GPT-4o & GPT-4.1: $25 per 1K tool calls o-series models: $10 per 1K tool calls For better performance, try web search with the o-series. If speed is critical, GPT-4o and GPT-4.1 remain the fastest options.
https://community.openai.com/t/deep-research-in-the-api-webhooks-and-web-search-with-o3/1299919
