The ‘deep research’ agent is designed to cater to the needs of those involved in knowledge-intensive work, such as finance, science, policy, and engineering. It can provide detailed, accurate, and trustworthy research. Moreover, it can assist in research for high-value purchases, like automobiles, appliances, etc. The research output, which can take 5 to 30 minutes, is fully documented with clear citations and a summary of its thought process.

This new agent comes after the January 23rd launch of ‘Operator’, another ChatGPT agent capable of performing tasks like grocery shopping and booking vacations online. The announcement follows the recent commotion caused by the arrival of a new AI model from China-based DeepSeek, which reportedly matched the performance of ChatGPT but at a much lower cost.

According to reports, there is an ongoing investigation by Microsoft and OpenAI into a potential breach of ChatGPT’s API data by a group associated with DeepSeek. OpenAI has claimed that their ‘deep research’ agent achieved a record score on the AI assessment called ‘Humanity’s Last Exam’, which features 3,000 expert-level questions on more than 100 topics.

The ‘deep research’ agent utilizes OpenAI’s latest ‘reasoning model’, the o3 model, which is designed for web browsing and data analysis. Despite its impressive performance, OpenAI cautions users that the agent may occasionally produce incorrect inferences or ‘hallucinate’ facts, and may have difficulty distinguishing between authoritative information and rumors.

In January, Google revealed plans to implement a similar feature, also named ‘Deep Research’, in its AI model Gemini by early 2025. OpenAI’s ‘deep research’ agent is currently available for its Pro plan subscribers at $200 per month, limited to 100 queries a month.

The post OpenAI’s Latest ChatGPT Agent Boasts ‘Deep Research’ Capabilities for Knowledge Intensive Work appeared first on Crypto Market Insights: Dive In with CryptoUpdate.io.