Perplexity Launches Deep Research Tool - Freemium Access

Perplexity Introduces Advanced AI Research Capability

Perplexity has recently joined the forefront of AI innovation by unveiling a sophisticated research tool, with the announcement made on Friday. This new feature is designed to enhance the depth and reliability of information gathering.

Competition in AI-Powered Research

Similar functionalities have already been released by other major players in the AI landscape. Google introduced a comparable feature for its Gemini AI platform in December. Following suit, OpenAI launched its own research agent earlier this month. Notably, all three companies have adopted the same designation for these tools: Deep Research.

Focus on Professional Applications

The primary objective of these tools is to deliver more comprehensive and well-sourced answers, catering to professional needs beyond the scope of typical consumer chatbots. In a blog post, Perplexity emphasized that Deep Research is “proficient in a wide array of specialized tasks—spanning areas like finance, marketing, and product investigation.”

Availability and Usage

Currently, Perplexity Deep Research is accessible via the web interface. The company has indicated that integration into its Mac, iOS, and Android applications will follow shortly. Users can activate the feature by selecting “Deep Research” from a drop-down menu when submitting their inquiries. This initiates the creation of a detailed report, exportable as a PDF or shareable as a Perplexity Page.

How Deep Research Operates

The report generation process involves an iterative approach. Perplexity explained that Deep Research “repeatedly searches, analyzes documents, and determines the subsequent steps, refining its research strategy as it gains further understanding of the topic.” This mimics the methodology a human researcher might employ when exploring a new subject.

Performance Benchmarks

The company showcased the tool’s performance on Humanity’s Last Exam, a challenging AI benchmark featuring expert-level questions across diverse academic disciplines. Perplexity’s Deep Research achieved a score of 21.1% on this test. This result surpassed the performance of models like Gemini Thinking (6.2%), Grok-2 (3.8%), and OpenAI’s GPT-4o (3.3%), though it fell slightly short of OpenAI’s Deep Research (26.6%).

Pricing and Speed Advantages

A key differentiator is the pricing structure. While OpenAI’s Deep Research currently requires a $200 monthly Pro subscription, Perplexity’s Deep Research is offered for free. Non-subscribers receive a limited number of queries daily, while subscribers enjoy unlimited access. Furthermore, Perplexity’s Deep Research appears to be faster, typically completing tasks within three minutes, compared to the 5 to 30 minutes required by OpenAI’s offering.

Comparative Overview

Perplexity provided a comparative analysis of the various deep research products, outlining differences in technology, pricing, and performance across various applications and subject areas, with links to relevant articles. The key distinctions were summarized as follows:

• Perplexity AI is distinguished by its speed and ease of access, making it ideal for casual research.
• OpenAI excels in providing analytical depth, catering to the needs of enterprise-level users.
• Google offers the most seamless integration with existing productivity tools and workflows.

Potential Limitations and Considerations

Despite the advancements, concerns remain regarding the potential impact of these tools on the research process. As highlighted by The Economist, potential drawbacks include limitations in “creative” data interpretation, a reliance on readily available sources, and the risk that delegating research entirely to an AI assistant could “diminish opportunities for original thought.”