ChatGPT 'Deep Research' agent: OpenAI's answer to DeepSeek's rise?
The OpenAI Deep Research agent demonstrates greater accuracy than DeepSeek R1 in humanity's last test.
AI is the current hot topic in tech. After significant attention on the Chinese startup DeepSeek, all the big AI companies are making their presence known. ChatGPT, a leader in AI, faces tough competition from DeepSeek. Now, ChatGPT is responding to the DeepSeek buzz by introducing its new AI Agent, "Deep Research." This gives users more control and the ability to get research reports on complex topics in just minutes.
OpenAI CEO Sam Altman announced the feature on X (formerly Twitter), describing it as "like a superpower; experts on demand!" This suggests OpenAI is targeting professionals across various industries who need research support but may lack the time or resources to conduct extensive investigations themselves. The promise of rapid results is central to Altman's statement.
This new model is particularly relevant for professionals in fields like finance, science, policy, and engineering, where high-quality research is essential. Users can initiate a report by clicking the "Deep Research" button alongside the search button before submitting a prompt to ChatGPT. Within 5 to 30 minutes, the AI agent delivers a comprehensive summary and citations, clearly indicating the sources of information rather than simply fabricating it.
 |
Image: OpenAI |
This move by OpenAI is both strategic and timely. In December, Google launched a deep research feature for Gemini, albeit limited to Gemini Advanced users. Google also recently unveiled Gemini Flash 2.0, boasting improved speed and accuracy. Mirroring its competitors, OpenAI's Deep Research Agent is initially available to Pro users, with a limit of 100 queries per month. Access for Enterprise Team and Plus users will follow shortly.
OpenAI has presented data suggesting that ChatGPT's Access model is notably faster and more accurate than the competition. They claim a 26.6% accuracy rate on the Humanity's Last Exam test, significantly outperforming DeepSeek's R1 at 9.4%, and even surpassing its own GPT-4 at 3.3% and OpenAI's o3-mini (high) at 13.0%.
The Future of AI Dominance
Whether ChatGPT will maintain its AI dominance remains to be
seen. The introduction of Deep Research
intensifies the ongoing competition.
Ultimately, user satisfaction and the ability to effectively solve
real-world problems will be the deciding factors in this race. The AI landscape continues to evolve rapidly,
and the next breakthrough is always just around the corner.