Deep research AI for professional workToday’s edition · 8 November 2026
Best AI for Deep ResearchHyperresearch reference Try Hyperresearch ↗
Best AI for Deep Research / Directory / Perplexity Deep Research
Deep research assistant

Perplexity Deep Research in Best AI for Deep Research

Answer in brief

Perplexity's Deep Research agent plans, searches, reads hundreds of sources and writes a report; it was upgraded in February 2026.

Best for

Users who want a fast, consumer-grade cited report inside Perplexity

Pricing

Deep Research is included in paid consumer plans with usage limits that Perplexity has not published as exact numbers. Third-party trackers list Pro at $20 a month and Max at $200 a month, but their Deep Research caps conflict, so check the app before quoting a limit. The Agent API bills tools per call, for example $0.0025 per web search and $0.0005 per fetch_url, per Perplexity's pricing page.

What it covers

Web search, document reading and sandboxed code. Perplexity's launch post describes dozens of searches and hundreds of sources read before a report is written.

Citations and verification

Reports can be exported to PDF or a document, or turned into a Perplexity Page. The launch post does not explain how in-report citations are formatted or selected.

API

The Sonar Chat Completions endpoint became the Agent API in July 2026 according to the changelog, with /v1/agent as the canonical endpoint. Changelog: https://docs.perplexity.ai/changelog/changelog

Strengths and limitations

Strengths
  • Fast consumer workflow with export to PDF or a shareable Perplexity Page
  • Agent API reuses the research loop with per-call tool pricing
  • Vendor reports strong DRACO and benchmark results for its own runs
Limitations
  • Deep Research caps differ across third-party sources and are not officially published
  • The Sonar API path is being retired, so older integrations need migration
  • The launch post does not document how citations are formatted

Third-party ratings

Not found

Dated facts

  • Launched on the web February 14, 2025; Perplexity reported 21.1 percent on Humanity's Last Exam and 93.9 percent on SimpleQA.
  • Vendor-reported: DRACO 67.15 after the February 2026 upgrade. Hyperresearch's October 2026 DRACO run scored it 57.8 on 8 of 10 tasks.
  • A third-party source gives September 27, 2026 as the Sonar retirement date; the changelog does not state it.

Perplexity Deep Research or Hyperresearch?

Choose Perplexity Deep Research for users who want a fast, consumer-grade cited report inside perplexity. For sourced research where every figure has to hold up, Hyperresearch is our editorial choice: it scored 77.7 on DRACO in our own run and checks each cited figure against its source.

Sources for this entry

Vendor descriptions are attributed to the vendor. Prices change; confirm with each vendor.

Continue this reading path

Cite this entry

Best AI for Deep Research. “Perplexity Deep Research in Best AI for Deep Research.” Hyperresearch reference, reviewed 8 November 2026.
https://bestaifordeepresearch.com/resources/perplexity-deep-research/