Perhaps you’ve been there – staring at stacks of open windows on your computer screen and typing in yet another query in a search box. You’re on a research odyssey that sometimes seems close to a breakthrough, getting close to the crucial nuggets of information you seek. But it can also seem like you’re going in circles. You step outside for a much-needed break, clutching some liquid caffeine, and you fill your lungs with fresh air. When you return to your computer, you stare at the screen and wonder where you left off.
Research can be overwhelming, but there are a growing number of AI tools out there to help, the latest being Perplexity AI’s Deep Research which is like a virtual research assistant.
Perplexity, based in San Francisco, launched its deep research tool on February 14, 2025 aimed at saving users from such a scenario.
“Today we’re launching Deep Research to save you hours of time by conducting in-depth research and analysis on your behalf. When you ask a Deep Research question, Perplexity performs dozens of searches, reads hundreds of sources, and reasons through the material to autonomously deliver a comprehensive report. It excels at a range of expert-level tasks – from finance and marketing to product research – and attains high benchmarks on Humanity’s Last Exam.”
Perplexity is the latest AI chatbot to offer a deep research tool, joining the likes of OpenAI’s ChatGPT Deep Research and Google Deep Research. Perplexity’s Deep Research tool will be free, but for a limited number of research queries – five, per day. Perplexity Pro subscribers get 500 Deep Research queries for $20 a month.
Perplexity’s Deep Research is available on the web and will soon be rolling out to iOS, Android, and Mac, according to the company.
“Deep Research takes question answering to the next level by spending 2-4 minutes doing the work it would take a human expert many hours to perform,” it said in the press release.
There are dozens of versions of AI tools that are geared towards saving humans time when it comes to research. For journalists, for example, there is Google’s Pinpoint. Pinpoint is not a deep research tool in the sense of what has recently emerged, but Pinpoint nonetheless can take a deep dive into reams of documents and pull together a concise analysis in mere minutes – work that would take humans months to complete.
“Using Pinpoint you can upload and search hundreds of thousands of documents, images, emails, hand-written notes, and audio files for specific words or phrases, locations, organizations, and people.”
The deep research tools mark an evolution. Whereas Pinpoint and other such tools can collect, sort and analyze immense volumes of information, the recently emerged deep research tools can do that and provide reader-ready reports.
Perplexity Deep Research is embedded with search and coding capabilities with the ability to read documents, reason about what to do next and refine research plans as it learns more about the subject matter. This is comparable to what a human would do as they learn more about a subject through their research, and are guided by their acquisition of new knowledge.
But where a human has to spend hours – if not days – to combine volumes of information into a concise report, Perplexity Deep Research can do it in minutes, and then format the report into a PDF or document or convert it into a Perplexity Page to share with friends or colleagues, the company says.
Perplexity’s launch comes just two weeks after the ChatGPT Deep Research launch. OpenAI stated, “Deep research is built for people who do intensive knowledge work in areas like finance, science, policy, and engineering and need thorough, precise, and reliable research. It can be equally useful for discerning shoppers looking for hyper-personalized recommendations on purchases that typically require careful research, like cars, appliances, and furniture. Every output is fully documented, with clear citations and a summary of its thinking, making it easy to reference and verify the information. It is particularly effective at finding niche, non-intuitive information that would require browsing numerous websites. Deep research frees up valuable time by allowing you to offload and expedite complex, time-intensive web research with just one query.”
OpenAI said its version may take anywhere from 5 to 30 minutes to complete its work. “In the meantime, you can step away or work on other tasks—you’ll get a notification once the research is complete. The final output arrives as a report within the chat – in the next few weeks, we will also be adding embedded images, data visualizations, and other analytic outputs in these reports for additional clarity and context.”
Humanity’s Last Exam (HLE) is an evaluation of Large Language Models (LLMs) by researchers at the Center for AI Safety, with collaboration from Scale AI. The evaluation relied on approximately 2,700 test questions posed to AI systems across a broad range of subject matter, including, but not limited to, math, physics, biology, computer science, chemistry, engineering and humanities.
HLE included questions from approximately 1,000 subject experts, including professors, researchers and graduate degree holders, affiliated with over 500 institutions from across 50 countries. The results of the evaluation were released last month.
Perplexity said its Deep Research earned a 21.1% accuracy score which is higher than most of the others tested, but below OpenAI’s Deep Research score of 26%.
In HLE’s synopsis of its testing, it stated, “As AI systems approach human expert performance in many domains, precise measurement of their capabilities and limitations is essential for informing research, governance, and the broader public.”
