OpenAI Launches Deep Research for Advanced Web Summarization
OpenAI has launched Deep Research, an advanced system designed to browse the web, summarize content, and answer questions based on these summaries. This release has been met with considerable enthusiasm, especially due to its high performance on the challenging General AI Assistants benchmark (GAIA). Deep Research integrates a large language model (LLM) with an internal agentic framework, allowing it to utilize tools like web search efficiently. It achieves up to 67% accuracy on one-shot questions and nearly 48% on complex level 3 inquiries, which require sophisticated reasoning over multiple steps. However, OpenAI has not fully disclosed the details of the underlying framework, leading the AI community, with Hugging Face at the forefront, to attempt to replicate the system and offer an open-source version. This initiative seeks to improve agentic frameworks through a CodeAgent approach, simplifying action expression and reducing operational steps, among other benefits. Ongoing community efforts continue to enhance AI assistance capabilities.