Perplexity review

The full review

Perplexity is the AI tool that has most fundamentally changed how a particular kind of user — researchers, journalists, knowledge workers, anyone who used to live inside Google's search results — does their work. The product is an AI search engine: ask a question in natural language, get a synthesized answer with inline citations to the sources the model drew from. For users whose work requires "citation-backed answer to a specific question, fast," Perplexity has done what no other AI tool has — made search feel like a step backward.

The Pro tier at twenty dollars a month is where Perplexity earns its place in a stacked AI subscription budget. The free tier is good and probably enough for casual users, but Pro unlocks the best models — GPT-5, Claude 4 Sonnet, and Perplexity's in-house Sonar variants — along with deeper search modes, file uploads, and the threading that makes long research sessions productive. For users who use Perplexity many times per day, Pro is obvious value.

The strongest case for Perplexity is research depth versus speed. Asked a question that would take an hour of Google searching, link-following, and synthesis, Perplexity returns a synthesized answer with citations in under a minute — and the citations are linked, scannable, and let the user verify or dig deeper as needed. For knowledge workers, journalists, analysts, and anyone whose job involves "answer this question with sources," Perplexity is genuinely transformative.

The Spaces feature, where users can curate sets of sources and run repeated searches against them, is the under-discussed productivity multiplier. Researchers building a dossier on a topic can point Perplexity at a specific set of academic papers, news archives, or company filings and get search results scoped to their corpus. That changes the depth of analysis that becomes practical.

Real weaknesses. Perplexity is not a coding tool. The code-eval scores reflect this — the model can write code, but the product surface is not built for the iterative, file-aware workflows that define modern coding tools. Engineers will reach for Cursor, Claude Code, or Copilot for actual programming. Perplexity is also not the right tool for creative writing, image generation, or voice interaction — it is a focused product, and that focus is its strength rather than a limitation, but new users sometimes expect the breadth ChatGPT delivers.

Recommendation: Perplexity Pro at twenty dollars a month is essential for users whose work involves research, current-events questions, or citation-heavy output. For users whose AI use is mostly writing, coding, or general assistant tasks, Perplexity is a complementary subscription rather than a replacement for ChatGPT or Claude. The combination of Perplexity for research plus Claude or ChatGPT for writing-and-thinking is one of the most common power-user stacks we've seen, and it works.