Perplexity AI has launched "Deep Research," a new feature designed to deliver comprehensive research reports by conducting in-depth analysis within minutes. The tool, announced on Friday, aims to provide expert-level insights across multiple domains, including finance, marketing, and product research.
The launch of Deep Research follows similar advancements from major AI companies. Google introduced a comparable feature for its Gemini AI platform in December 2024, and OpenAI followed suit earlier this month. Notably, all three companies have opted to name their tools "Deep Research."
Perplexity's Deep Research mode operates by performing extensive searches, reviewing numerous sources, and synthesising information into structured reports. The company claims that the tool can accomplish within minutes what would typically require hours of manual research. Users can access this feature through the Perplexity website by selecting "Deep Research" from the mode selector before submitting a query.
According to Perplexity, Deep Research has demonstrated strong performance on Humanity’s Last Exam, a benchmarking test that evaluates AI models on expert-level questions. The tool scored 21.1%, outperforming competitors such as Google’s Gemini Thinking (6.2%), Grok-2 (3.8%), and OpenAI’s GPT-4o (3.3%). However, OpenAI’s own Deep Research tool scored higher at 26.6%.
While OpenAI’s Deep Research feature currently requires a Pro subscription costing $200 per month, Perplexity has opted for a more accessible pricing model. The tool is available for free, though non-subscribers are limited to five queries per day, whereas Pro users can access up to 500 queries daily. Additionally, Perplexity’s Deep Research is reported to be significantly faster, completing most tasks in under three minutes, compared to OpenAI’s 5 to 30 minutes.
Perplexity claims the following differentiators:
Users can try out the new Deep Research tool on the web, and the rollout on iOS, Android, and Mac will happen soon.