Perplexity AI

Best for: Real-time questions with cited answers

The core Perplexity experience: ask any question and receive a synthesised, cited answer drawn from real-time web search. Every claim links to its source, making it instantly verifiable — a significant advantage over AI tools that generate plausible-sounding but unverifiable responses. The free tier uses Perplexity's own Sonar model.

Best for: Advanced research with choice of AI model

Pro unlocks access to GPT-4o, Claude, Gemini, and Sonar Huge as the reasoning engine — allowing users to choose the model best suited to each research task. Higher query limits, file upload support, and image generation are included. The right tier for professionals using Perplexity as a primary research tool.

Best for: Comprehensive autonomous research reports

Deep Research runs autonomously — searching, reading, and synthesising dozens of sources before producing a comprehensive, structured report with full citations. It performs the equivalent of several hours of human desk research in three to five minutes. Ideal for market analysis, competitor intelligence, and due diligence starting points.

Best for: Search-optimised reasoning and retrieval

Sonar is Perplexity's own large language model, specifically trained for web-grounded reasoning — combining search retrieval with language generation in a single optimised system. It is faster and more cost-efficient than using third-party models for search-heavy tasks, and powers the default Perplexity experience.

Best for: Maximum capability for complex research tasks

Sonar Huge is Perplexity's largest and most capable proprietary model — available via the API and Enterprise Pro plans. It delivers deeper reasoning and synthesis than standard Sonar, and is suited to complex multi-source research tasks that require sustained analytical depth alongside real-time retrieval.

Best for: Shared AI research environments for teams

Spaces are collaborative research environments where teams work together within a shared AI context. Members can upload shared documents, set a shared system prompt, and build on each other's research threads — making it the closest Perplexity gets to a true team research platform. Ideal for ongoing market monitoring or client research programmes.

Best for: Saving and organising research threads

Collections allow users to save, organise, and revisit research conversations grouped by topic or project. Useful for maintaining a running research library across an ongoing project — preserving context and sources without needing to repeat searches.

Best for: Turning research into shareable reports

Pages converts Perplexity research into formatted, shareable web-based reports with a polished layout. Research compiled in Perplexity can be exported as a Pages document, making it suitable for client delivery, team briefing documents, or internal knowledge bases without additional formatting work.

Best for: Private, secure research for organisations

Perplexity Enterprise Pro adds SSO, admin controls, private search (conversations not used for training), usage analytics, and unlimited queries for all team members. Designed for organisations where research involves sensitive or confidential information that cannot be shared with a general AI platform.

Best for: Real-time web search in your applications

The Sonar API provides developer access to Perplexity's real-time search and synthesis capabilities — enabling applications to query the live web through an AI layer. Used for building research tools, automated briefing systems, and products where up-to-date information is a core requirement.