Perplexity
What Is Perplexity?
Perplexity is an answer engine: you ask, and it returns a short, cited answer instead of a list of blue links. It combines large language models with live web search and attaches numbered sources you can open. In 2026 it has grown past search, adding multi-step Research reports and an agent called Computer, but its core value is still fast research with a visible paper trail.
It is not a citation engine for academic rigor. It synthesizes across sources faster than a browser can, but citations point to pages, not verified truth; open the source before relying on any claim.
Key Features
- Cited answers: numbered source links on every response, including Free.
- Live web search: up-to-date results, with focus modes for academic papers, Reddit, YouTube, or a chosen domain.
- Model orchestration: queries are routed to a suitable model; paid plans allow a preferred model per query.
- Pro Search and Research: multi-step modes that plan, search repeatedly, and synthesize findings.
- Computer: an agent on paid plans for longer tasks such as reports and app building, metered with credits.
Pricing: Free, Pro, and Max (September 2026)
The Free plan includes practically unlimited Quick searches with citations, enough for light daily research. The limits sit elsewhere: the current help-center table lists about 3 Pro Searches per day and roughly one Research query per month, plus limited file uploads, no model selection, and no premium data sources. Expect these numbers to drift; an older help page said 5 Pro Searches per day.
Pro costs $20 per month, or $17 per month billed annually. It raises Pro Search to weekly limits sized for average use, adds a monthly allowance of Research and file-and-app creation queries, unlocks advanced models with per-query selection (currently including Claude Sonnet 5, Gemini 3.1 Pro, GPT-5.6, and Sonar 2), lifts file uploads to about 50 per Space, and adds image and video generation, premium data sources such as PitchBook and Statista, and priority support. New subscribers get 4,000 one-time bonus credits for Computer. Max costs $200 per month ($167 annual) with 10,000 monthly credits and the highest limits. Enterprise plans are quoted through sales; verified students and educators get a discounted Education Pro.
| Plan | Price | What you get |
|---|---|---|
| Free | $0 | Unlimited Quick search, ~3 Pro Searches/day, ~1 Research query/month |
| Pro | $20/mo ($17 annual) | Weekly Pro Search limits, model selection, higher uploads, Computer access |
| Max | $200/mo | 10,000 credits/month, frontier models, highest limits |
Pro Search, Research, and Computer in Practice
A Quick search does one retrieval pass and summarizes it. Pro Search plans the question, runs several searches, and merges the results, which matters for comparisons and ambiguous queries. Research (formerly Deep Research) runs longer investigations and outputs a structured report; the free tier gets about one per month. The old Labs feature has effectively folded into file-and-app creation and Computer. Computer is metered by credits: 100 credits equal $1, light tasks use roughly 15 to 70 credits, and tasks pause rather than fail when credits run out.
Perplexity vs ChatGPT Search and Google
ChatGPT with search enabled is the closest rival: stronger as a general assistant and drafting tool, with less consistent citations. Perplexity is citation-first on every answer, including the free tier. Google still has the deepest index and leaves the judging to you, at the cost of doing the synthesis yourself. A practical split: Perplexity for cited quick answers, ChatGPT for writing and iteration, Google when you need to inspect primary results directly.
API Pricing
Perplexity publishes API rates in its official docs (checked September 2026). The Search API costs $5 per 1,000 requests with no token charges. Sonar models bill per token: Sonar at $1 per 1M input and output tokens, Sonar Pro at $3 and $15, Sonar Reasoning Pro at $2 and $8, plus request fees that rise with search context size. The Agent API bills at each model's published rates plus tool fees, such as $0.0025 per web search. Sonar chat completions retire on September 27, 2026, replaced by the Agent API.
Limitations and Trade-offs
Answer quality depends on the sources it retrieves, and a citation is not a guarantee of authority. Free-tier limits interrupt heavy research quickly, Pro Search allowances are weekly bands rather than hard numbers, and the product is built for answers, not deep interactive writing or long-term memory. Perplexity has revised allowances downward more than once, so treat every number here as a snapshot.
Who Should Use Perplexity?
It fits researchers, analysts, students, journalists, and anyone who wants fast, source-backed answers without opening ten tabs; the free tier is enough for light use. It fits poorly if you mainly need a writing or coding partner, verbatim citation of primary documents, or stable quotas. For document-grounded Q&A over your own files, compare Notion AI.