Sonar Deep Research

Exhaustive research with expert-level insights

A powerful research model capable of conducting exhaustive searches across hundreds of sources, synthesizing expert-level insights, and generating detailed reports with comprehensive analysis.

Input Tokens.

Per 1M

$2

Output Tokens

Per 1M

$8

Citation Tokens

Per 1M

$2

Search Queries

Per 1K

$5

Reasoning Tokens

Per 1M

$3

Features

Arrow

Deep research / Reasoning model

Search

Exhaustive research across hundreds of sources

Length

128K context length

Speed

Expert-level subject analysis

Citation

Detailed report generation

Lock

No training on customer data

Real World Use Cases

Academic Research

Academic research and comprehensive reports

Market Analysis

Market analysis and competitive intelligence

Due Diligence

Due diligence and investigative research

curl --request POST \
  --url https://api.perplexity.ai/chat/completions \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '{
  "model": "sonar-deep-research",
  "messages": [
    {"role": "user", "content": "Provide an in-depth analysis of the impact of AI on global job markets over the next decade."}
  ],
  "max_tokens": 500
}'
Sample Response Metadata
{"id": "8b6b7969-a4ff-488d-a8b9-b3fca3698cce", 
"model": "sonar-deep-research", 
"created": 1747668848, 
"usage": {"prompt_tokens": 19, "completion_tokens": 498, "total_tokens": 517, "citation_tokens": 10175, "num_search_queries": 48, "reasoning_tokens": 95305}, 
"citations": ["https://www.indiatoday.in/technology/news/story/ai-could-shake-up-job-market-by-2030-mckinsey-reveals-list-of-sectors-that-will-be-impacted-2547147-2024-06-02", "https://etradeforall.org/news/future-of-jobs-report-2025-the-jobs-of-the-future-and-the-skills-you-need-to-get-them", "https://www.unleash.ai/future-of-work/pwc-hr-become-an-ai-disruptor-by-being-skills-first/", "https://vorecol.com/blogs/blog-automation-and-the-impact-on-the-workforce-10891", "https://edisonandblack.com/pages/over-97-million-jobs-set-to-be-created-by-ai.html", "https://www.ferolabs.com/insights/post/can-ai-create-manufacturing-jobs", "https://barrychaiken.com/archives/barrypchaiken/2025/04/1005", "https://www.mckinsey.com/featured-insights/future-of-work/jobs-lost-jobs-gained-what-the-future-of-work-will-mean-for-jobs-skills-and-wages", "https://www.marketingaiinstitute.com/blog/ai-impact-on-jobs", "https://www.ilo.org/meetings-and-events/generative-ai-and-jobs-global-analysis-potential-effects-job-quantity-and", "https://www.ilo.org/resource/article/how-reskilling-ai-could-unlock-new-and-better-jobs", "https://opentools.ai/news/ai-impact-unctad-predicts-automation-could-shake-40percent-of-jobs-and-widen-global-disparities", "https://www.mckinsey.com/mgi/our-research/generative-ai-and-the-future-of-work-in-america", "https://cfo.economictimes.indiatimes.com/news/85-of-2030-jobs-yet-to-be-invented-amid-rapid-industry-transformation-dell-report/120375527", "https://news.outsourceaccelerator.com/?p=75453", "https://accesspartnership.com/understanding-how-ai-impacts-jobs-and-skills-in-asean/", "https://mexicobusiness.news/cloudanddata/news/ai-may-impact-38-latin-america-works", "https://www.oecd.org/en/publications/artificial-intelligence-and-wage-inequality_bf98a45c-en.html", "https://www.goldmansachs.com/insights/articles/ai-may-start-to-boost-us-gdp-in-2027"], 
"object": "chat.completion", 
"choices": [{"index": 0, "finish_reason": "length", 
"message": {"role": "assistant", "content": "<think>\nOkay, so the user wants an in-depth analysis of how AI will impact global job markets over the next decade. Let me start by breaking down the query. They're looking for a comprehensive report, so I need to cover various aspects like job creation, displacement, sector-specific impacts, skills required, and regional differences. \n\nFirst, I should check the search results provided. There are 20 sources here, from McKinsey, World Economic Forum, ILO, Goldman Sachs, etc. Let me skim through them to identify key points. \n\nSource  mentions AI creating 69 million jobs by 2028. Source [1] says McKinsey predicts 12 million occupational transitions by 2030. Source [2] from the World Economic Forum talks about 170 million new jobs this decade. Source [3] from PwC compares AI to the industrial revolution, suggesting job creation might offset losses. Source [4] has conflicting numbers: 800 million jobs displaced but 555 million new ones created. Source [5] says 97 million new jobs by 2025. Source [6] focuses on manufacturing, saying AI can fill 2 million jobs by 2030. Source [7] is about healthcare jobs being both displaced and created. Source [8] from McKinsey again, automation affecting 15% of hours worked globally. Source [9] Brookings says 30% of workers could have half their tasks affected. Source [10] ILO study says AI will augment rather than replace jobs. Source [11] emphasizes reskilling. Source [12] UNCTAD report says 40% of jobs impacted, widening disparities. Source [13] McKinsey on US job shifts. Source [14] Dell report claims 85% of 2030 jobs not invented yet. Source [15] Africa's tech outsourcing could see 40% task disruption. Source [16] ASEAN skills changing 72% by 2030. Source [17] Latin America 38% jobs impacted. Source [18] OECD on wage inequality. Source [19] Goldman Sachs predicts US GDP boost from AI.\n\nSo, the main themes seem to be: job displacement vs creation, sectoral shifts (healthcare, STEM, manufacturing, administrative), regional differences (developed vs developing nations), skills transformation, and economic impacts like productivity and inequality. \n\nI need to structure the report into sections. Let's outline possible sections:\n\n1"}, 
"delta": {"role": "assistant", "content": ""}}]}

Advanced Features

Control the computational effort dedicated to each query with the reasoning_effort parameter. This allows you to balance between speed and thoroughness while managing costs by directly impacting the amount of reasoning tokens consumed.Options:
  • "low": Faster, simpler answers with reduced token usage
  • "medium": Balanced approach (default)
  • "high": Deeper, more thorough responses with increased token usage
curl --request POST \
  --url https://api.perplexity.ai/chat/completions \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '{
    "model": "sonar-deep-research",
    "messages": [{"role": "user", "content": "What should I know before markets open today?"}],
    "reasoning_effort": "low"
  }'