Perplexity AI Launches Free 'Deep Research' Feature, Challenging Competitors in AI-Powered Research Tools

What is Perplexity Deep Research, and how do you use it?

Besides being better than Google for search, Perplexity, the artificial intelligence (AI) company, wants to be an expert on any subject with its new Deep Research feature. This cutting-edge tool, launched by Perplexity AI in February 2025, combines autonomous reasoning with rapid processing to deliver exhaustive reports on specialized topics. Also: Perplexity is the AI tool Gemini wishes it could be According to Perplexity, "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." The company claims at its core that Perplexity Deep Research employs a proprietary framework called test time compute (TTC) expansion, which enables the systematic exploration of complex topics. Unlike conventional search engines that retrieve static results, the TTC architecture mimics human cognitive processes by iteratively refining its understanding through analysis cycles. The system begins by dissecting the query into subcomponents, then autonomously performs dozens of web searches, evaluates hundreds of sources, and synthesizes findings through probabilistic reasoning models. Also: What is DeepSeek AI? Is it safe? Here's everything you need to know This layered approach allows the AI to reconcile contradictory information, identify emerging patterns, and prioritize authoritative sources -- a capability proven by its 21.1% score on the rigorous "Humanity's Last Exam" AI benchmark. That may sound lousy but, by comparison, GPT-4o scored 3.1%, and DeepSeek-R1 came in with 8.5%. Perplexity describes this reasoning as "refining its research plan as it learns more about the subject areas. This is similar to how a human might research a new topic, refining one's understanding throughout the process." Alternatively, Ken Huang, CEO of DistributedApps.ai and VP of research at CSA GCR, describes TTC as a "model [that] takes input data and applies its learned parameters to produce an output. For neural networks, this involves forward propagation through the network layers using matrix multiplications and activation functions." I don't know about you, but that doesn't sound much like how I reason out the answer to a question. Also: US sets AI safety aside in favor of 'AI dominance' However, Perplexity Deep Research uses parallelized data ingestion and hierarchical summarization techniques to deliver expert-level reports in two to four minutes. A human researcher might spend hours on the same request. Early tests suggest that Perplexity's latest AI tool is faster than Google's Deep Research for Gemini and OpenAI's Deep Research. I do wonder, what with all this talk about how revolutionary AI is for business, why all three companies ended up describing their serious research functionality as "Deep Research". True creativity doesn't appear to be any AI's strong suit. So, Perplexity is very fast, but can it deliver the goods? The answer is "sort of". A TechRadar review suggested Perplexity hallucinated quite a lot. In my tests, I asked the program to dig deeply into three subjects I'm an expert in and that few others are. Those subjects in full: the history and influence of the Commercial Internet Exchange (CIX); the role of Sir Édouard Percy Cranwill Girouard in the East Africa Protectorate; and the history of x86 Unix desktop distributions. Also: Perplexity lets you try DeepSeek R1 without the security risk Perplexity delivered a useful abstract for all three topics but in no way an expert-level report. For that depth, you need to hire me. As before, I prefer Perplexity's output over other AI chatbots because its inline citations make it easy to double-check its answers. In addition, while Perplexity didn't make major blunders, it made enough minor ones, so there's no question in my mind that you can't just turn in a Perplexity report and expect it to pass muster. No, you still need to check its answers. Welcome to the State of AI in 2025. Of course, you're unlikely to need answers to any of those questions. Perplexity claims it's good for finance, marketing, and product research, so I gave it a question near and dear to my heart: "Tell me how to make a commercially successful Linux desktop." The answers the AI gave me sounded good, but all too often, when it came to the fine details, they were wrong. For example, the AI said companies and users want long-term stability from their desktop operating systems. That's true. After all, there's a reason why, according to Statcounter, 69% of desktop computers worldwide are still running Windows 10. However, in the same paragraph that mentions this truism, the report states Ubuntu Linux only has a six-month lifecycle. Ah, wrong. You can now run Ubuntu Linux with support for up to a dozen years. I think that's enough long-term stability for anyone. Also: I tested DeepSeek's R1 and V3 coding skills - and we're not all doomed (yet) In short, you still need real experts to double-check Perplexity's homework even for its target subjects. Still, Perplexity Deep Research disrupts the premium pricing trend in advanced AI research tools by offering free access to Deep Research, albeit with daily query limits. Non-subscribers receive five free daily queries -- sufficient for casual research needs -- while Pro subscribers, $20/month, get 500 daily queries. That price is much cheaper than OpenAI's Deep Research, available to ChatGPT Pro subscribers at $200/month. On the other hand, Google Gemini Advanced, which comes with its Deep Research, has numerous other features and costs $20 a month. Anyone can try Perplexity Deep Research. The tool requires minimal technical expertise: Advanced users can enhance output quality through the following options: Pro subscribers gain additional features, such as custom template creation and application programming interface access for batch processing. Native Android and iOS apps are scheduled for a second-quarter 2025 rollout, but, for now, you must use the web interface. Also: I bought an iPhone 16 for its AI features, but I haven't used them even once - here's why So, is Perplexity Deep Research worth trying? Check it out for yourself and see. You can do enough with the AI for free to get a good idea if it's helpful. Perplexity is still the most useful tool for me when I use it as a replacement for Google, which is why I pay for a subscription. Used with caution, I think many people will find it useful as a starting place for serious research projects.

Perplexity AI's Deep Research Tool Is Almost as Good as OpenAI's, and It's Free

After OpenAI, Perplexity AI is joining the "deep research" bandwagon. And it's doing it in a fairly interesting way. Following in the footsteps of DeepSeek's "reasoning" model, Perplexity is the first major AI provider that's offering a Deep Research feature for free users, too. By comparison, OpenAI's Deep Research feature is only available in the $200/month Pro subscription. Deep Research is an upcoming AI feature that takes a bit of time, but performs dozens of related searches, goes over hundreds of resources, and uses a reasoning model to logic out each prompt in a step-by-step process. You can get similar results from tools like Copilot's "Think Deeper" feature, sure, but what sets Deep Research apart is that it puts all of the info it's collected together into a comprehensive, white paper style report. Deep Research is free for all logged-in users, though you're limited to just five queries a day. But if you're paying for Perplexity Pro (which costs $20/month or $200/year), you get up to 500 queries a month. OpenAI's Deep Research feature is limited to 100 queries a month for now. Although, it's not an apples-to-apples comparison. ChatGPT's Deep Research feature takes a lot of time. Up to 20 minutes. It asks follow up questions, shows all the complex steps involved in its process, and spits out a very long report at then end. Perplexity's Deep Research feature is kind of a lite version of that. You get a response in 2-4 minutes, so your results will naturally rely more in searching the web and data collation than deep interpretation on the part of the AI. OpenAI is using its upcoming o3 reasoning model for Deep Research, but Perplexity hasn't mentioned any details about the model it's using. In Humanity's Last Exam, a commonly used AI benchmark consisting of over 3,000 questions across a number of topics, Perplexity's Deep Research scored a 21.1% on accuracy, which is much higher than DeepSeek R1 (8.6%), and Gemini (7.2%). OpenAI's Deep Research still has the lead with a 26.6% completion score, but a silver medal is respectable here given the tool's much lower barrier to entry. So, how does this change how you might use Perplexity? So far, AI chatbots have been all about multiple prompts. You ask a question and prompt again and again to get to detailed answers. But with Deep Research, you can ask a single question and be done with it. The more specific and verbose your prompt, the better the bot's research and report will be, but the AI can now give you pages of info in response to much less prompting. And once your report is generated, you can download it as a PDF. Over on its blog, Perplexity has highlighted multiple examples of the kind of difference Deep Research can make. Where the default model might givee an overview with bullet-point answers, Deep Research will instead come back with multiple paragraphs, detailed reports, and more expanded formatting.

Perplexity AI's Deep Research feature is available now -- here's how to try it for free

Following ChatGPT and Google Gemini deep research advancements, Perplexity AI has unveiled its latest feature, "Deep Research," designed to give users the tools they need to do extensive research in just minutes, providing comprehensive answers to their complex queries. Deep Research is particularly adept at handling expert-level queries across various domains, including finance, marketing, and product research. It has demonstrated strong performance on industry benchmarks, such as scoring 21.1% on "Humanity's Last Exam," outperforming several other models. The extensive research feature operates by leveraging advanced AI algorithms to sift through vast amounts of data, delivering detailed responses that would traditionally require hours of manual research. The introduction of Deep Research comes at a pivotal time for Perplexity AI. Founded in 2022 by former OpenAI researcher Aravind Srinivas, the company has rapidly ascended in the tech industry, attracting significant investments from notable figures such as Jeff Bezos and tech giant Nvidia. As of December 2024, Perplexity secured a $500 million funding round, elevating its valuation to $9 billion. Recently, Perplexity put in a bid to partner with TikTok U.S. Deep Research is currently accessible to all users. Those with free accounts are allotted up to five queries per day, while Pro users benefit from an expanded limit of 500 queries daily. Initially only available on the web, Perplexity AI plans to extend this feature to iOS, Android, and Mac platforms. To utilize Deep Research, users should go to the Perplexity AI website and select "Deep Research" from the mode options in the search box before submitting their query. This streamlined process enables users to generate in-depth research reports on any topic efficiently. Perplexity has seen rapid growth despite legal challenges from major media organizations alleging unauthorized use of copyrighted material. In October 2024, News Corp subsidiaries, including The Wall Street Journal and The New York Post, filed a lawsuit accusing Perplexity of "massive freeriding" on their content. The lawsuit claims that Perplexity's AI-driven search engine reproduces substantial portions of their articles without permission or compensation. In response to these allegations, Perplexity AI has initiated efforts to collaborate with publishers. The company announced revenue-sharing agreements with media outlets such as Time and Fortune, aiming to compensate publishers when their content is cited in AI-generated answers. This move reflects a broader industry trend, as AI companies seek to balance technological advancement with respect for intellectual property rights. Despite these challenges, Perplexity AI continues to innovate. The company operates on a freemium model, offering both free and paid enterprise versions of its services. The free model utilizes the company's standalone language model based on GPT-3.5 with browsing capabilities, providing personalized search results with summarized information and inline citations. The paid "Pro" version offers additional features, including access to an API, the ability to search both internal files and web content, and the use of advanced language models such as GPT-4 and Claude 3.5 within the Perplexity site. The launch of Deep Research underscores Perplexity AI's commitment to enhancing the efficiency and depth of information retrieval while maintaining their position as a leader in AI. By enabling users to obtain thorough answers rapidly, this feature has the potential to transform workflows across various industries. As the company navigates the complexities of innovation and intellectual property rights, it remains at the forefront of the evolving landscape of AI-powered search technologies.

Is This the End of Traditional Research? Perplexity AI Thinks So

Perplexity AI's Deep Research can handle expert-level tasks across finance, marketing, product research, and other domains. Inspired by OpenAI, and Google, Perplexity AI has launched Deep Research, a tool for autonomously conducting in-depth research and analysis. The feature performs multiple searches, reviews hundreds of sources, and compiles findings into comprehensive reports. It is free for everyone -- up to 5 queries per day for non-subscribers and 500 queries per day for Pro users. According to Perplexity, Deep Research can handle expert-level tasks across finance, marketing, product research, and other domains. "We believe everyone should have access to powerful research tools," the company stated. "Excited to introduce the Perplexity Deep Research Agent: available for free to all users. Paid users only need to pay $20/mo to access an expert level researcher on any topic for 500 daily queries, and need to wait less than three minutes for getting a full research report," said Perplexity AI co-founder and CEO Aravind Srinivas in a post on X. On the other hand, OpenAI's deep research is available as part of the ChatGPT Pro plan, which costs $200 per month. This plan includes a limit of 100 queries per month. The feature is initially available on the web and will soon be rolled out to iOS, Android, and Mac. Users can access it by selecting "Deep Research" from the mode selector before submitting a query. Perplexity claims its Deep Research model scores 21.1% on Humanity's Last Exam, outperforming Gemini Thinking, o3-mini, o1, DeepSeek-R1, and other top models. It also records 93.9% accuracy on the SimpleQA benchmark, surpassing other leading AI models. The company positions Deep Research as a tool for various applications, including financial analysis, market research, and personal consulting. "Deep Research takes question answering to the next level," Perplexity stated. The AI system is designed to function similarly to human researchers, refining its understanding as it gathers more data. Deep Research operates in three phases, including iterative research, report writing, and exportability. The AI searches for information, evaluates sources, refines its approach, and synthesises findings into a report. Users can then export reports as PDFs, documents, or Perplexity Pages for sharing. The company has also optimised Deep Research for speed. Notably, the company recently announced that its in-house model, Sonar, will be available to all Pro users on the platform. Now, users with the Perplexity Pro plan can make Sonar the default model by changing settings. Sonar is built on top of Meta's open-source Llama 3.3 70B. It is powered by Cerebras Inference, which claims to be the world's fastest AI inference engine. The model can produce 1200 tokens per second.

Perplexity launches its own freemium 'deep research' product | TechCrunch

Perplexity has become the latest AI company to release an in-depth research tool, with a new feature announced Friday. Google unveiled a similar feature for its Gemini AI platform in December. Then OpenAI launched its own research agent earlier this month. All three companies even have given the feature the same name: Deep Research. The goal is to provide more in-depth answers with real citations for more professional use cases, compared to what you'd get from a consumer chatbot. In a blog post announcing Deep Research, Perplexity wrote that the feature "excels at a range of expert-level tasks -- from finance and marketing to product research." Perplexity Deep Research is currently available on the web, and the company said it will soon be added to its Mac, iOS, and Android apps. To use it, you just select "Deep Research" from a drop-down menu when you submit your query in Perplexity, which will then create a detailed report that can be exported as a PDF or shared as a Perplexity Page. To create this report, 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," supposedly "similar to how a human might research a new topic." The company also highlighted its performance on Humanity's Last Exam, an AI benchmarking test with expert-level questions in a variety of academic fields. Perplexity said its Deep Research tool scored 21.1% on the test, easily beating most other models, such as Gemini Thinking (6.2%), Grok-2 (3.8%), and OpenAI's GPT-4o (3.3%) -- but not quite matching OpenAI's Deep Research (26.6%). But while you currently need a $200-per-month Pro subscription to use OpenAI's Deep Research (the company plans to expand to other subscription tiers), Perplexity's Deep Research is available for free -- non-subscribers get an unspecified-but-limited number of queries per day, while paying subscribers get unlimited queries. Perplexity's Deep Research also seems to perform more quickly, completing most tasks in under three minutes compared to 5 to 30 minutes for OpenAI Deep Research. Asked to compare the various deep research products, Perplexity offered an overview of the different technologies, pricing models, and performance in different use cases and subject matters, with links to articles about each feature). It summarized the differences as follows: While it's too early to know how these tools will affect everyday and professional research as they become more popular, The Economist recently highlighted shortcomings to OpenAI's Deep Research that likely apply here too: not just limitations to its "creativity" in interpreting data and a tendency to rely on sources that are "easily available," but a larger risk that "outsourcing all your research to a supergenius assistant" could "reduce the number of opportunities to have your best ideas."

Perplexity Launches a Free 'Deep Research' AI Tool

Perplexity is the latest AI chatbot to get a dedicated "Deep Research" tool. ChatGPT got its version of the tool earlier this month, and Google's Gemini got one in December. Much like its rivals, Perplexity's Deep Research tool can fetch information from hundreds of sources "to autonomously deliver a comprehensive report." It excels at a range of expert-level tasks in areas like finance, marketing, and technology, Perplexity says in a blog post. To get started, users can go to perplexity.ai, select Deep Research from the drop-down menu in the chat box, and drop their query. The AI tool will then review sources, refine its research plan, and develop a report in 2-4 minutes. Users can download the prepared report as a PDF/document or convert it into a shareable Perplexity page. On Humanity's Last Exam, a benchmark that evaluates the reasoning abilities of AI models, the accuracy of Perplexity's Deep Research tool was topped only by OpenAI's Deep Research (26.6%). DeepSeek R1, Gemini, and Grok-2 lag behind at 8.6%, 7.2%, and 3.9%, respectively. At launch, Perplexity's Deep Research tool will be free for all logged-in web users. It will roll out to iOS, Android, and Mac. However, only Perplexity Pro subscribers will get unlimited queries. Others will have to contend with a limited number of answers per day, Perplexity said. In comparison, Gemini's Deep Research tool requires a Gemini Advanced subscription that comes with a Google One AI plan ($19.99 per month), and ChatGPT's tool requires a ChatGPT Pro subscription ($200 per month). However, OpenAI CEO Sam Altman said recently that the company plans to open more advanced tiers to free users this year.

Perplexity Launches Deep Research to Rival OpenAI and Google

Those on the free tier will get a limited number of queries per day Perplexity introduced a Deep Research function on its platform last week. The tool can analyse complex questions better and provide comprehensive reports by reasoning through the topics. The new feature is being added to the model picker and is currently available to all users. The company claims that its responses are generated faster than the eponymous OpenAI tool, with an average output time of three minutes. Notably, Perplexity recently introduced an artificial intelligence (AI) assistant for Android devices with computer vision and web search features. After Google and OpenAI, now Perplexity has become the latest Silicon Valley AI company to introduce its Deep Research product. Based on the description of the feature in the blog post, the feature likely uses a test time compute expansion technique to allow the AI model to spend more time on a query. This way, the AI can delve deeper into the problem, second guess its response and verify, as well as evaluate alternate theories. Perplexity is rolling out Deep Research to all users. Pro subscribers will get an unlimited number of queries using the tool, while those on the free tier will get an unspecified but limited number of queries per day. Currently, the feature is only available on Perplexity's web client, but the company stated that it will also be rolled out to the iOS, Android, and Mac apps. To use Deep Research, Perplexity users can tap on the mode selected underneath the text box and find the option at the bottom of the list. Once selected, the AI will automatically begin answering queries using the Deep Research function. The company has claimed that a response takes about three minutes to generate on average, which is much faster than the five to 30-minute period OpenAI's tool takes. The tool uses search and coding capabilities to conduct iterative research, reading documents and adjusting its approach based on the information gathered. After evaluating the source materials, it compiles the findings into a structured report. Users can export the report as a PDF or document or share it as a Perplexity Page with others. Notably, the company claimed that Deep Research scored 21.1 percent on Humanity's Last Exam in an internal test, outsourcing Gemini Thinking, OpenAI's o1 and o3-mini, and DeepSeek-R1.

Perplexity Launches Deep Research for AI-Powered Expert Analysis

Available for free with query limits; unlimited access for Pro subscribers. AI search company Perplexity has launched Deep Research, a new feature designed to automate expert-level research and analysis. The tool rapidly conducts in-depth research by performing dozens of searches, reading hundreds of sources, and synthesizing findings into comprehensive reports -- all within minutes. "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," Perplexity said in a blog post on February 14. Also Read: OpenAI Launches Deep Research: AI Agent for In-Depth Web Analysis According to the company, Deep Research functions like a human researcher, iteratively refining its understanding of a topic. It systematically searches for relevant information, processes complex materials, and compiles clear, structured reports. Users can export findings as PDFs or Perplexity Pages for easy sharing. "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," the company said. "We built Deep Research to empower everyone to conduct expert-level analysis across a range of complex subject matters," Perplexity said, adding that Deep Research is designed for tasks in finance, marketing, technology, health, current affairs, and travel planning. It achieves a 21.1 percent accuracy on Humanity's Last Exam, outperforming AI models like Gemini Thinking, o3-mini, o1, DeepSeek-R1, and many other leading models, and scores 93.9 percent on the SimpleQA benchmark. Most reports are generated in under three minutes, the company said, adding that it is working to make them even faster in the future. Also Read: Perplexity Acquires Carbon, Expands Publisher Program with 15 New Media Partners Deep Research is free for all users, with daily query limits for non-subscribers. Pro subscribers get unlimited Deep Research queries. The feature is currently available on web, with rollouts planned for iOS, Android, and Mac. To try Deep Research, visit perplexity.ai and select it from the mode menu selector in the search box before submitting the query.

Perplexity one-ups Gemini and ChatGPT with a fantastic AI freebie

What if you tell an AI chatbot to search the web, look up a certain kind of source, and then create a detailed report based on all the information it has gleaned? Well, Gemini can do it, for $20 a month. Or $200 each month, if you prefer ChatGPT. Perplexity will do it for free. A few times each day, that is. Perplexity is calling its latest tool, Deep Research. Just like OpenAI. And Google Gemini before it. Recommended Videos So, what's the fuss all about? Think of it as a research assistant, who takes the drudgery of looking up sources, taking notes, and then handing over a well-prepared report. Performing a task like that takes a lot of computing power, and as a result, it is expensive. Perplexity will offer an unlimited number of Deep Research queries to its subscribers, but free users can enjoy "a limited number of answers per day." "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," says the company. Deep Research by Perplexity is currently limited to the web, but will soon arrive on Android and iOS mobile apps, as well. To launch a Deep Research query, users will have to select the namesake option from the model selection dropdown, right by the text field. Once the research work is over, users will be able to export it directly as a document or PDF file. There's also an option to convert it into an online Perplexity Page, which is shareable with anyone as a web link. In Gemini's case, you can directly import it to Google Docs, as well. The company claims Perplexity Deep Research is better than Google's Gemini model with Thinking capabilities, OpenAI's o3-mini and o1 models, and the buzzy new DeepSeek-R1 model. It only ranks below OpenAI Deep Research on the Humanity's Last Exam benchmark. I've extensively used Gemini Deep Research for looking through scientific papers and archives, and it does a fantastic job of pulling up the relevant information and simplifying it. It is also particularly effective at pulling up notices from government agencies and handles legal updates quite well. OpenAI CEO, Sam Altman, also confirmed a few days ago that the company will offer two Deep Research queries to free users per month. For those paying a $20 monthly fee for ChatGPT Plus, they will get 10 Deep Research shots each month.

Out-analyzing analysts: OpenAI's Deep Research pairs reasoning LLMs with agentic RAG to automate work -- and replace jobs

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More Enterprise companies need to take note of OpenAI's Deep Research. It provides a powerful product based on new capabilities, and is so good that it could put a lot of people out of jobs. Deep Research is on the bleeding edge of a growing trend: integrating large language models (LLMs) with search engines and other tools to greatly expand their capabilities. (Just as this article was being reported, for example, Elon Musk's xAI unveiled Grok 3, which claims similar capabilities, including a Deep Search product. However, it's too early to assess Grok 3's real-world performance, since most subscribers haven't actually gotten their hands on it yet.) OpenAI's Deep Research, released on February 3, requires a Pro account with OpenAI, costing $200 per month, and is currently available only to U.S. users. So far, this restriction may have limited early feedback from the global developer community, which is typically quick to dissect new AI advancements. With Deep Research mode, users can ask OpenAI's leading o3 model any question. The result? A report often superior to what human analysts produce, delivered faster and at a fraction of the cost. How Deep Research works While Deep Research has been widely discussed, its broader implications have yet to fully register. Initial reactions praised its impressive research capabilities, despite its occasional hallucinations in its citations. There was the guy who said he used it to help his wife who had breast cancer. It provided deeper analysis than what her oncologists provided on how radiation therapy was the right course of action, he said. The consensus, summarized by Wharton AI professor Ethan Mollick, is that its advantages far outweigh occasional inaccuracies, as fact-checking takes less time than what the AI saves overall. This is something I agree with, based on my own usage. Financial institutions are already exploring applications. BNY Mellon, for instance, sees potential in using Deep Research for credit risk assessments. Its impact will extend across industries, from healthcare to retail, manufacturing, and supply chain management -- virtually any field that relies on knowledge work. A smarter research agent Unlike traditional AI models that attempt one-shot answers, Deep Research first asks clarifying questions. It might ask four or more questions to make sure it understands exactly what you want. It then develops a structured research plan, conducts multiple searches, revises its plan based on new insights, and iterates in a loop until it compiles a comprehensive, well-formatted report. This can take between a few minutes and half an hour. Reports range from 1,500 to 20,000 words, and typically include citations from 15 to 30 sources with exact URLs, at least according to my usage over the past week and a half. The technology behind Deep Research: reasoning LLMs and agentic RAG Deep Research does this by merging two technologies in a way we haven't seen before in a mass-market product. Reasoning LLMs: The first is OpenAI's cutting-edge model, o3, which leads in logical reasoning and extended chain-of-thought processes. When it was announced in December 2024, o3 scored an unprecedented 87.5% on the super-difficult ARC-AGI benchmark designed to test novel problem-solving abilities. What's interesting is that o3 hasn't been released as a standalone model for developers to use. Indeed, OpenAI's CEO Sam Altman announced last week that the model instead would be wrapped into a "unified intelligence" system, which would unite models with agentic tools like search, coding agents and more. Deep Research is an example of such a product. And while competitors like DeepSeek-R1 have approached o3's capabilities (one of the reasons why there was so much excitement a few weeks ago), OpenAI is still widely considered to be slightly ahead. Agentic RAG: The second, agentic RAG, is a technology that has been around for about a year now. It uses agents to autonomously seek out information and context from other sources, including searching the internet. This can include other tool-calling agents to find non-web information via APIs; coding agents that can complete complex sequences more efficiently; and database searches. Initially, OpenAI's Deep Research is primarily searching the open web, but company leaders have suggested it would be able to search more sources over time. OpenAI's competitive edge (and its limits) While these technologies are not entirely new, OpenAI's refinements -- enabled by things like its jump-start on working on these technologies, massive funding, and its closed-source development model -- have taken Deep Research to a new level. It can work behind closed doors, and leverage feedback from the more than 300 million active users of OpenAI's popular ChatGPT product. OpenAI has led in research in these areas, for example in how to do verification step by step to get better results. And it has clearly implemented search in an interesting way, perhaps borrowing from Microsoft's Bing and other technologies. While it is still hallucinating some results from its searches, it's doing so less than competitors, perhaps in part because the underlying o3 model itself has set an industry low for these hallucinations at 8%. And there are ways to reduce mistakes still further, by using mechanisms like confidence thresholds, citation requirements and other sophisticated credibility checks. At the same time, there are limits to OpenAI's lead and capabilities. Within two days of Deep Research's launch, HuggingFace introduced an open-source AI research agent called Open Deep Research that got results that weren't too far off of OpenAI's -- similarly merging leading models and freely available agentic capabilities. There are few moats. Open-source competitors like DeepSeek appear set to stay close in the area of reasoning models, and Microsoft's Magentic-One offers a framework for most of OpenAI's agentic capabilities, to name just two more examples. Furthermore, Deep Research has limitations. The product is really efficient at researching obscure information that can be found on the web. But in areas where there is not much online and where domain expertise is largely private -- whether in peoples' heads or in private databases -- it doesn't work at all. So this isn't going to threaten the jobs of high-end hedge-fund researchers, for example, who are paid to go talk with real experts in an industry to find out otherwise very hard-to-obtain information, as Ben Thompson argued in a recent post (see graphic below). In most cases, OpenAI's Deep Research is going to affect lower-skilled analyst jobs. The most intelligent product yet When you merge top-tier reasoning with agentic retrieval, it's not really surprising that you get such a powerful product. OpenAI's Deep Research achieved 26.6% on Humanity's Last Exam, arguably the best benchmark for intelligence. This is a relatively new AI benchmark designed to be the most difficult for any AI model to complete, covering 3,000 questions across 100 different subjects. On this benchmark, OpenAI's Deep Research significantly outperforms Perplexity's Deep Research (20.5%) and earlier models like o3-mini (13%) and DeepSeek-R1 (9.4%) that weren't hooked up with agentic RAG. But early reviews suggest OpenAI leads in both quality and depth. Google's Deep Research has yet to be tested against this benchmark, but early reviews suggest OpenAI leads in both quality and depth. How it's different: the first mass-market AI that could displace jobs What's different with this product is its potential to eliminate jobs. Sam Witteveen, cofounder of Red Dragon and a developer of AI agents, observed in a deep-dive video discussion with me that a lot of people are going to say: "Holy crap, I can get these reports for $200 that I could get from some top-4 consulting company that would cost me $20,000." This, he said, is going to cause some real changes, including likely putting people out of jobs. Which brings me back to my interview last week with Sarthak Pattanaik, head of engineering and AI at BNY Mellon, a major U.S. bank. To be sure, Pattanaik didn't say anything about the product's ramifications for actual job counts at his bank. That's going to be a particularly sensitive topic that any enterprise is probably going to shy away from addressing publicly. But he said he could see OpenAI's Deep Research being used for credit underwriting reports and other "topline" activities, and having significant impact on a variety of jobs: "Now that doesn't impact every job, but that does impact a set of jobs around strategy [and] research, like comparison vendor management, comparison of product A versus product B." He added: "So I think everything which is more on system two thinking -- more exploratory, where it may not have a right answer, because the right answer can be mounted once you have that scenario definition -- I think that's an opportunity." A historical perspective: job loss and job creation Technological revolutions have historically displaced workers in the short term while creating new industries in the long run. From automobiles replacing horse-drawn carriages to computers automating clerical work, job markets evolve. New opportunities created by the disruptive technologies tend to spawn new hiring. Companies that fail to embrace these advances will fall behind their competitors. OpenAI's Altman acknowledged the link, even if indirect, between Deep Research and labor. At the AI Summit in Paris last week, he was asked about his vision for artificial general intelligence (AGI), or the stage at which AI can perform pretty much any task that a human can. As he answered, his first reference was to Deep Research: "It's a model I think is capable of doing like a low-single-digit percentage of all the tasks in the economy in the world right now, which is a crazy statement, and a year ago I don't think something that people thought is going to be coming." (See minute three of this video). He continued: "For 50 cents of compute, you can do like $500 or $5,000 of work. Companies are implementing that to just be way more efficient." The takeaway: a new era for knowledge work Deep Research represents a watershed moment for AI in knowledge-based industries. By integrating cutting-edge reasoning with autonomous research capabilities, OpenAI has created a tool that is smarter, faster and significantly more cost-effective than human analysts. The implications are vast, from financial services to healthcare to enterprise decision-making. Organizations that leverage this technology effectively will gain a significant competitive edge. Those that ignore it do so at their peril. For a deeper discussion on how OpenAI's Deep Research works, and how it is reshaping knowledge work, check out my in-depth conversation with Sam Witteveen in our latest video:

Perplexity just made AI research crazy cheap -- what that means for the industry

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More Perplexity shattered the artificial intelligence market's status quo today by launching Deep Research, a tool that generates comprehensive research reports in minutes and opens advanced AI capabilities to users at a fraction of typical enterprise costs. "Thankful for open source! We're going to keep making this faster and cheaper," wrote Perplexity CEO Aravind Srinivas in a post on X.com. "Knowledge should be universally accessible and useful. Not kept behind obscenely expensive subscription plans that benefit the corporates, but not in the interests of humanity!" Perplexity Deep Research is redefining AI pricing -- can enterprise AI survive? The launch exposes a painful truth in AI pricing: expensive enterprise subscriptions may be unnecessary. While Anthropic and OpenAI charge thousands monthly for their services, Perplexity offers five free queries daily to all users. Pro subscribers pay $20 monthly for 500 daily queries and faster processing -- a price point that could force larger AI companies to explain why their services cost up to 100 times more. Companies have been significantly increasing their AI investments, with enterprise AI spending expected to rise by 5.7% in 2025, despite overall IT budget increases of less than 2%. Some businesses are planning to increase their AI spending by 10% or more, with an average increase of $3.4 million dedicated to AI initiatives. These investments now look questionable as Perplexity delivers similar capabilities at consumer prices. How Perplexity Deep Research is outperforming Google and OpenAI Deep Research's technical achievements suggest expensive AI services may be overpriced rather than superior. The system scored 93.9% accuracy on the SimpleQA benchmark and reached 20.5% on Humanity's Last Exam, outperforming Google's Gemini Thinking and other leading models. "Deep Research on Perplexity completes most tasks in under 3 minutes," the company announced, highlighting its ability to perform dozens of searches and analyze hundreds of sources simultaneously. The tool combines web search, coding capabilities, and reasoning functions to refine research iteratively -- mimicking expert human researchers but at machine speed. Why Perplexity's affordable AI is breaking down barriers to advanced technology The implications stretch beyond pricing. Enterprise AI has created a digital divide between well-funded companies and everyone else. Small businesses, researchers, and professionals who couldn't afford thousand-dollar subscriptions were effectively locked out of advanced AI capabilities. Perplexity's approach changes this calculation. The tool handles complex tasks from financial analysis and market research to technical documentation and healthcare insights. Users can export findings as PDFs or share them through Perplexity's platform, potentially replacing expensive research subscriptions and specialized tools. The company plans to expand Deep Research to iOS, Android, and Mac platforms, which could accelerate adoption among users who previously viewed AI tools as out of reach. This broad access may prove more valuable than any technical breakthrough -- finally putting advanced AI capabilities in the hands of users who need them most. For technical decision makers, this shift demands attention. Companies paying premium prices for AI services should examine whether those investments deliver value beyond what Perplexity now offers at a fraction of the cost. The answer may reshape how organizations approach AI spending in 2025 and beyond. While Perplexity's competitors scramble to justify their premium pricing, thousands of users are already testing Deep Research's capabilities. Their verdict might matter more than any benchmark: in AI's new reality, the best technology isn't the one that costs the most -- it's the one people can actually use.