Perplexity

Perplexity — Answer Engine + Agentic Browser + Cloud Computer (100M MAU, $30B Valuation Talk)

Perplexity (perplexity.ai / perplx.ai, San Francisco, founded Aug 2022 by Aravind Srinivas, Denis Yarats, Johnny Ho, Andy Konwinski; ex-Google/OpenAI engineers) active 2026-08, no discontinuation. FY2026 annualized revenue crossed $750M (from <$250M Jan 2026), 100M+ MAU, Nvidia in talks to invest at >$30B valuation​ (Sept 2025 round was $20B; $1.5B+ raised total incl. Bezos/SoftBank/IVP/NEA), IPO target 2028. Product now three layers: answer engine (citations) → Perplexity Computer (cloud agent) → Comet browser. Category ai-productivity-data; pipeline_stage none. Killer differentiator = every answer is source-verifiable by design​ (citations are the product, not an add-on) + multi-model orchestration (~20 models: Claude Opus 5/GPT-5.6/Gemini/Sonar/Nano Banana/Veo 3.1/Grok, route-per-task, no single model >25% by Dec 2025)​ + Deep Research (40 sources in ~4 min, 90%+ citation accuracy) + Labs/Pages + Spaces + Brain memory + Computer (Firecracker-isolated VM) + Comet (Chromium, agent reads live tabs); not Google Gemini Deep Research (2M ctx, deeper narrative, but 30× quota advantage at $20 and weaker inline citations), not ChatGPT Deep Research (broader but no citation-grounded engine heritage), not You.com (weaker agent/browser), not traditional search (loses “explore by links” discovery). 2026 trap: quota engineering is the whole story​ — Free = unlimited quick search but 5 Pro searches/day; Pro $20 = Deep Research caps repeatedly cut (Pro Search 600→200/week in early 2026, now 300+/day); Max $200 for unlimited frontier + Computer + 10,000 credits; Max tier introduced Apr 2026; Comet went fully free Mar 18 2026​ (was Max-only $200, now $0 browser, only Comet Plus $5 add-on or Pro/Max include it); citations still hallucinate on high-stakes (source-check everything); open-web only (no Bloomberg/IEEE gated); Comet agent mis-clicks dense pages + early-2026 security research: weaker phishing block + MCP-exploit class; no Linux first-party; scientific PDFs trail Gemini; publisher deals (Forbes/WSJ/FT/Le Monde via Comet Plus) created “content theft” lawsuits + opt-out backlash.

1. One-line positioning

Perplexity = the research stack that starts where search ends: type a question → Sonar model fires dozens of real-time searches → reads full source docs → synthesizes with numbered inline citations you click-verify​ (reasoning chain visible) → follow-up threads keep context → Free 5 Pro/day, Pro $20 unlocks Deep Research (40-source reports ~4 min, 90%+ accuracy), model picker (GPT-5.6 / Claude Sonnet 5 / Sonar 2), file/image upload, image gen, Labs (code/data/asset generation), Pages (shareable formatted reports), Spaces (team curated research desks with custom instructions) → Brain (Max, 15 langs, self-improving wiki memory, +10pp memory benchmark, -6% dissatisfaction) → Perplexity Computer​ (Feb 25 2026, Firecracker microVM, orchestrates 20 models, Slack @computer + web/mobile, enterprise Snowflake/Salesforce/SharePoint/HubSpot/Datadog + SSO/SCIM/audit/SOC2/zero-retention) → Comet browser​ (Chromium, address bar = answer engine, sidecar reads current tab, agent fills forms/compares pricing/cross-tab summary, voice, ad/tracker blocker, Finance Hub, Model Council on Max; free Win/Mac/iOS/Android since Mar 18 2026, Comet Plus $5 publisher content). Fits ai-productivity-data; pipeline_stage none (query→search→synthesize→verify→report→agent-act). For journalists/investors/analysts/policy researchers needing auditable real-time intelligence, it’s unmatched; for long-horizon academic PDF synthesis, Gemini 2.5 Pro deeper; for creative/open chat, ChatGPT/Claude.

2. Core features

3. Pricing (2026-08, USD)

pricing_type = freemium-plus-per-seat-plus-usage-credits-plus-api-plus-enterprise-quote; starting_price = $0; free_tier = "yes; unlimited quick/Sonar searches, 5 Pro searches/day, basic models + search history, no files/upload/image-gen/Labs/Pages/Spaces/Brain/Computer, Comet browser fully free since Mar 18 2026"; pricing_note = "perplexity-sf-aug-2022-srinivas-yarats-ho-konwinski-active-2026-no-shutdown; max-tier-introduced-apr-2026-200; computer-launched-feb-25-2026; comet-free-march-18-2026-was-max-only; pro-search-quota-cut-600-to-200-weekly-early-2026-now-300-plus-day; annual-pro-200-16.67-mo-17pct-off; comet-plus-5-publisher-content-free-pro-max; api-sonar-5-per-1m-input-2-per-1m-cached; enterprise-org-wide-credit-pool-group-caps-rbac-scim-jul-2026; 100m-mau-750m-arr-nvidia-talks-30b-valuation-ipo-2028"

4. Access Type

access_type: web-app-plus-mobile-plus-browser-plus-api-plus-slack

access_display: 🌐 perplexity.ai (search/spaces/labs/pages/computer) · 📱 iOS/Android · 🖥️ Comet browser (Win/Mac/iOS/Android, Chromium, no Linux first-party) · 🔌 API (Sonar) · 💬 Slack @computer (originally internal Slack bot) · connectors: 100+ (Salesforce/Snowflake/Datadog/HubSpot/SharePoint + MCP custom) · no desktop Electron separate from Comet

5. Reviews (2026)

Praise: citations are structural, auditable​ (every claim clickable — wins peer review / supervisor challenge), real-time web = same-day event coverage no other deep-research matches, Deep Research 4 min vs 2 hrs, multi-model = never stuck on one vendor’s bad day, Labs turns research into polished doc/app, Spaces is a private research desk, Brain remembers projects across sessions, Computer + Slack @computer collapses analyst/ops workflows, Comet free + Chromium = zero switching cost​ (bookmarks/extensions carry over), Finance Hub + live tickers, Model Council side-by-side on Max, enterprise RBAC/group credit caps + zero-retention finally procurement-ready.

Gripes: quota cuts eroded trust​ (Pro Search 600→200/week early 2026, Deep Research caps repeatedly revised; buy annual only after monthly test), citations still fabricate on high-stakes claims (verify sources manually), open-web only​ (Bloomberg/IEEE/terminal data out of reach — keep those subs), Comet agent mis-clicks dense/booking flows (supervise financial/irreversible), early-2026 security: weak phishing block + MCP code-injection exploit​ = not for sensitive tabs yet, no Linux build, scientific literature synthesis trails Gemini 2.5 Pro (2M ctx, narrative depth), publisher program (Forbes/WSJ/FT) sparked copyright lawsuits + opt-out backlash, Creative/long-form prose weaker than ChatGPT/Claude, free-tier 5 Pro/day throttles real work, Max $200 is a cliff for casuals.

6. Best for / Not for

7. Competitors