Perplexity has entered the growing contest to make AI tools more useful for serious research. Its new Deep Research feature is designed to produce more detailed answers with citations, moving beyond the faster, lighter responses people expect from consumer chatbots.
The launch puts Perplexity alongside Google and OpenAI, which have introduced similarly named Deep Research products for their own AI platforms. The difference Perplexity is emphasizing is access: its version is available for free, with limits for non-subscribers and unlimited use for paying subscribers.
A crowded race for deeper AI research
Perplexity is not the first major AI company to put a research-focused product into the market. Google unveiled a similar feature for its Gemini AI platform in December. OpenAI launched its own research agent earlier this month.
All three companies have used the same name: Deep Research. That shared label points to a common goal. These products are meant to give users longer, more source-backed outputs than a standard chatbot answer.
Perplexity described its feature as useful for expert-level work. In a blog post announcing Deep Research, the company wrote that the feature "excels at a range of expert-level tasks—from finance and marketing to product research."
That framing matters because the product is not positioned merely as a search shortcut. It is being presented as a tool for users who need a structured report, citations, and a process that can handle more complex questions.
How Perplexity Deep Research works
Perplexity Deep Research is currently available on the web. The company said it will soon be added to its Mac, iOS, and Android apps.
Using it is meant to be straightforward. A user selects "Deep Research" from a drop-down menu when submitting a query in Perplexity. The system then creates a detailed report.
That report can be exported as a PDF or shared as a Perplexity Page. Those output options suggest that Perplexity is thinking beyond one-off answers and toward research that can be saved, circulated, or used as a working document.
Perplexity said Deep Research "iteratively searches, reads documents, and reasons about what to do next, refining its research plan as it learns more about the subject areas." The company described that process as supposedly "similar to how a human might research a new topic."
The practical promise is a system that does not simply answer once and stop. Instead, it gathers information, adjusts its direction, and builds toward a final report. That is the distinction Perplexity is trying to draw between Deep Research and a more conventional chatbot response.
Performance claims and benchmark comparisons
Perplexity also highlighted how its Deep Research tool performed on Humanity’s Last Exam, an AI benchmarking test with expert-level questions across a variety of academic fields.
According to Perplexity, its Deep Research tool scored 21.1% on the test. That put it ahead of several named models in the source article, including Gemini Thinking at 6.2%, Grok-2 at 3.8%, and OpenAI’s GPT-4o at 3.3%.
It did not top OpenAI’s Deep Research, which was listed at 26.6%. Still, the benchmark comparison gives Perplexity a way to argue that its tool belongs in the same conversation as products from larger AI rivals.
The timing claims are also central to the pitch. Perplexity’s Deep Research appears to complete most tasks in under three minutes. OpenAI Deep Research is described as taking 5 to 30 minutes.
Those differences matter for how the tools might be used. A slower system may fit long-form investigation, while a faster one could be more practical for repeated professional questions throughout the day. The source does not prove how either will perform in every setting, but it shows how Perplexity is trying to define its advantage.
Pricing may be the sharpest difference
The clearest contrast is pricing. OpenAI’s Deep Research currently requires a $200-per-month Pro subscription, although the company plans to expand the feature to other subscription tiers.
Perplexity’s Deep Research is available for free. Non-subscribers get an unspecified-but-limited number of queries per day. Paying subscribers get unlimited queries.
That freemium model could make Perplexity’s tool more accessible to users who want to test deep research workflows without committing to a high monthly subscription. It also gives Perplexity a way to put its product in front of people who may not yet know whether they need this kind of AI research assistant.
The source does not specify the exact free query limit, so that remains an important unknown. For heavy users, the distinction between limited and unlimited access may be decisive. For occasional users, even a limited free version may be enough to understand the product’s value.
Why the human role still matters
The rise of Deep Research tools raises a bigger question: what happens when more research work is handed to AI systems?
The source article points to concerns recently highlighted by The Economist about OpenAI’s Deep Research that likely apply here too. Those concerns include limitations to its "creativity" in interpreting data and a tendency to rely on sources that are "easily available."
There is also a broader warning. The Economist raised the risk that "outsourcing all your research to a supergenius assistant" could "reduce the number of opportunities to have your best ideas."
That caution does not make tools like Perplexity Deep Research irrelevant. It does suggest they should be treated as research aids rather than replacements for judgment. A cited report can speed up information gathering, but the user still has to decide whether the framing is useful, whether the sources are enough, and what the findings actually mean.
Perplexity’s launch shows how quickly AI research products are becoming a competitive category. The next test is not only whether these systems can produce long reports, but whether professionals can use them without surrendering the thinking that makes research valuable in the first place.