OpenAI launches deep research function, generating analyst-level reports in 5 to 30 minutes

On February 4th, following the impact of DeepSeek, the release frequency of OpenAI's new product features has significantly increased. Following the release of the official o3-mini model three days ago and the free addition of the "inference" button in ChatGPT, on the morning of February 3rd Beijing time, OpenAI released a new ChatGPT feature called "Deep Research", which aims to generate analyst-level reports after 5 to 30 minutes of networking and deep thinking. Currently, Pro users can use this feature.

 

OpenAI CEO Sam Altman said that this is a service like "on-demand experts" that can complete tasks that previously took hours or days. However, OpenAI stated that this feature sometimes produces illusions or makes incorrect inferences in the answers, but the probability of errors is significantly lower than other models under existing ChatGPT.