Perplexity

Perplexity is an AI-powered search engine and answer product. You ask a question; it searches the web in real time, reads the top sources, synthesizes an answer, and shows you exactly which sources backed each claim. The output reads like a well-researched summary with footnotes — not like a chatbot guessing.

The product is purposely focused on search and research. It's not a general chatbot — for open-ended creative or coding work, you'll be better served by ChatGPT or Claude. Perplexity's superpower is that when you have a real question and you want a sourced answer, it gets you there faster than anything else.

The single most important thing about Perplexity is that every answer is sourced. Every claim has a footnote linking to the page that backs it. Every paragraph cites multiple sources. You can click any citation and verify the claim yourself, immediately.

This matters more than it sounds. Without citations, an AI answer is a guess you have to trust. With citations, an AI answer is a synthesis you can verify. For any work where being wrong has consequences — research, journalism, due diligence, technical documentation — Perplexity's citation discipline is transformative.

Through 2024 and 2025, Perplexity built out their own family of models — Sonar — optimized specifically for search-grounded answers. Sonar is the default on the free tier and the fastest option on Pro. Quality is genuinely strong: comparable to GPT-4 class models on the search-and-summarize tasks Perplexity is built for, dramatically faster, and significantly cheaper to run.

For 80% of search queries, Sonar is the right choice. For complex research questions where you want maximum reasoning, the Pro tier lets you switch to GPT, Claude, or Gemini.

Perplexity Pro at $20/month gives you a model picker that includes Sonar plus the latest production models from OpenAI (GPT-5), Anthropic (Claude Sonnet 4.6 and Opus 4.6), Google (Gemini 2.5 Pro), and a few others. You pick which model to route each query to.

This is genuinely useful. For quick lookups, Sonar. For nuanced reasoning, Claude. For coding, GPT-5. For multimodal, Gemini. Having all of them under one $20/month subscription is the same value proposition as AIBox — except Perplexity is search-first and AIBox is workspace-first.

Pages let you turn a Perplexity research session into a shareable document. Ask a series of questions, refine the answers, and Perplexity generates a structured page you can share with anyone or publish publicly. Useful for one-off briefs and quick reports.

Spaces are persistent collections — like project folders — where you can save searches, upload reference files, and organize ongoing research. Each Space has its own model and instructions, so you can set up (for example) a 'competitive intelligence' Space with specific instructions and feed it queries over time.

For anyone doing ongoing research on a topic, Spaces are the right organizational unit.

Comet is Perplexity's browser, launched in 2024 and refined into 2026. It builds Perplexity into the sidebar of every webpage, lets you query the page you're on, and integrates with the rest of Perplexity for cross-session research. Available to Pro and Enterprise Pro subscribers on macOS and Windows.

Is Comet better than Chrome with the Perplexity extension? Marginally. Most of Comet's value can be replicated with the extension on Chrome. But the integration is tighter and the experience cleaner — worth trying if you're already paying for Pro.

Free Perplexity is meaningfully useful — basic search, limited Pro Searches, Sonar access. Enough for casual research.

Perplexity Pro at $20/month gives 300+ Pro Searches per day, the model picker, unlimited Pages and Spaces, Comet browser access, and file analysis. The right tier for any heavy user.

Enterprise Pro at $40/user/month adds team Spaces, admin controls, no-training defaults, and expanded usage. The right tier for teams of 5+.

Max at $200/month gives you the highest-tier reasoning models, expanded research limits, and early access to new features. Overkill for most users.

It's not a great general chatbot. Try to use it for creative writing, brainstorming, or coding and you'll feel the friction. Perplexity is built around search; it does search well and other things less well.

No built-in image generation. You'll need DALL-E, Midjourney, Ideogram, or Imagen elsewhere.

Free tier is generous but the best features (Pro Search, model picker, Comet) are paywalled. The free tier is enough to evaluate but not enough to replace your other tools.

Perplexity is the best AI search product on the market. For any research-heavy work, Pro at $20/month pays for itself within a week. The combination of citation discipline, multi-model access, and persistent Spaces makes it the daily-driver tool for anyone whose work involves looking things up. Pair it with a general chatbot (ChatGPT, Claude, or AIBox) for everything that isn't search.