OpenAI's newest ChatGPT agent can do 'deep research' online ...
OpenAI has recently unveiled a new agent for its ChatGPT artificial intelligence tool, termed "deep research." This innovative feature empowers users to conduct in-depth online research to generate reports comparable to those produced by professional research analysts.
According to a blog post shared by OpenAI on February 2, deep research is specifically designed for individuals engaged in intensive knowledge work within fields such as finance, science, policy, and engineering. It aims to provide comprehensive, precise, and dependable research outcomes. The tool also proves beneficial for consumers looking to make informed decisions on significant purchases like cars or appliances. Reports generated by deep research are meticulously documented, including clear citations and a summary of the information presented. The process typically takes between 5 to 30 minutes to complete.
Enhancements and Comparison
This latest ChatGPT agent release comes shortly after the introduction of Operator on January 23. Operator, another ChatGPT agent, assists users in tasks like ordering groceries and arranging holiday tours using online resources. It also follows the emergence of a new AI model from DeepSeek, a China-based company, which caused a stir in the AI and US tech sectors due to its comparable performance to ChatGPT at a significantly lower cost.
OpenAI reported that its deep research agent achieved a remarkable milestone by scoring high on an AI assessment known as Humanity’s Last Exam. The evaluation consists of 3,000 expert-level questions spanning over 100 topics. The deep research agent obtained an accuracy rate of 26.6%, surpassing the scores of 9.4% for DeepSeek-R-1 and 3.3% for OpenAI's own GPT-4o model. This advanced agent is powered by the OpenAI o3 model, fine-tuned for web browsing and data analysis purposes. The o3 model, OpenAI's latest reasoning model, places a strong emphasis on fact-checking to mitigate inaccuracies and misinformation.
Challenges and Limitations
Despite its remarkable capabilities, OpenAI cautioned that the deep research agent may occasionally present inaccurate information or draw erroneous conclusions. The tool may struggle with differentiating between authoritative data and rumors, leading to potential inaccuracies in responses.
In a related development, Google announced plans to introduce a feature named "Deep Research" for its Gemini AI model in early 2025, mirroring OpenAI's recent innovation. OpenAI's deep research agent is currently accessible through its $200-a-month Pro plan, with a limitation of 100 queries per month.
For more information about DeepSeek R1, explore our top website in the world dedicated to all things investing. Additionally, discover custom scripts and ideas shared by our community members.
