PerplexityReasoningReleased 7 March 2025

Sonar Deep Research pricing, context and benchmark

Sonar Deep Research is a research-focused model designed for multi-step retrieval, synthesis, and reasoning across complex topics.

128K
Input context
115K
Output context
$2
Official input
$8
Official output
n/a
AA index

Everything that matters, in one table

Context, pricing, cache, modalities and tooling, from the model registry and Artificial Analysis.

ModelSonar Deep Research
Model IDperplexity/sonar-deep-research
ProviderPerplexity
Release date7 March 2025
StatusGenerally available
Input context128,000 tokens
Output context115,200 tokens
Official input pricing$2 per 1M tokens
Official output pricing$8 per 1M tokens
Cache read pricingNot published
Market input pricing$2 per 1M (cheapest route)
Market output pricing$8 per 1M (cheapest route)
Reasoning tokensBilled as output at $8 per 1M
Input modalitiestext
Output modalitiestext
Reasoning controlYes, effort levels exposed
Tool supportText generation only
TokenizerOther
Provider routes1 endpoint
Artificial Analysis indexNot published
SpeedNot published
Knowledge cutoffNot publicly stated

Input and output support

ModalityInputOutputBilled at
Text✓ yes✓ yes$2 in
Imagenononot supported
Audionononot supported
Videonononot supported
File (PDF)nononot supported

What is Sonar Deep Research?

Sonar Deep Research is a research-focused model designed for multi-step retrieval, synthesis, and reasoning across complex topics.

It autonomously searches, reads, and evaluates sources, refining its approach as it gathers...

Sonar Deep Research carries a 128K token window, uses the Other tokenizer and is served by Perplexity. On Hoodgen it sits in the Reasoning tier of the model menu next to its siblings, so you can switch mid conversation without losing the thread.

Release and what changed from the previous version

Released7 March 2025
PropertySonarSonar Deep ResearchChange
Released20252025Same release window
Context127K128KChanged
Max output114K115KChanged
Input price$1$2Higher
Output price$1$8Higher
Input modalitiestext, imagetextUnchanged
AA indexn/an/aNewer benchmark run

Features and capabilities

◈
Thinking control
Reasoning effort can be dialled per request, so one model covers deep analysis and fast chat.

Thinking and reasoning capabilities

Sonar Deep Research can think before answering and exposes the effort level per request. Reasoning tokens bill as output at $8 per million, and the trace can be returned with include_reasoning.

LevelFirst tokenCost per taskBest for
Measuredn/an/aArtificial Analysis measures the default reasoning configuration
Low effortFasterLowerChat, autocomplete, classification
High effortSlowerHigherDebugging, planning, multi step agents

What can you do with Sonar Deep Research?

Debugging and root cause
Multi file failures where the model has to hold a causal chain before answering.
Architecture and planning
Design docs, migration plans and trade off analysis over long context.
Data analysis
Spreadsheets and logs read in one pass, then summarised with the numbers quoted back.
Agentic coding
Repo wide changes that need plan, edit and verify loops.

Sonar Deep Research pricing

Official pricing

List price
$2 / 1M in
$8 / 1M out
  • 128K context
  • text input
Cheapest route
$2 / 1M in
$8 / 1M out
  • 1 route on OpenRouter
  • What Hoodgen routes to first
  • Same model, lower bill

Market pricing, every route

ProviderTierInputOutputCache readUptime 24h
Perplexitystandard$2$8n/an/a

API pricing and cost calculator

RatePer callTotalCredits
Cheapest route$0.0520$26.0024701
List price$0.0520$26.0024701

API and developer information

Model IDperplexity/sonar-deep-research
Endpointhttps://openrouter.ai/api/v1/chat/completions
Parametersfrequency_penalty · include_reasoning · max_tokens · presence_penalty · reasoning · temperature · top_k · top_p · web_search_options
StreamingYes, server sent events
Modality stringtext->text
Max prompt tokens128,000
Max completion tokens115,200
curl https://openrouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "perplexity/sonar-deep-research",
    "messages": [{ "role": "user", "content": "Summarise this document." }],
    "max_tokens": 1024
  }'

Model limitations

Text only outputIt cannot generate images, audio or video. Pair it with an image model on Hoodgen when you need visuals.
Output cap115,200 tokens per request. Long deliverables have to be chunked.
Benchmarks not publishedArtificial Analysis does not publish an Intelligence Index for this release, so no benchmark score is claimed here.
Knowledge cutoffThe provider does not state a fixed cutoff for every release. Verify very recent events against search.

Sonar Deep Research vs other Perplexity models

ModelContextInput /1MOutput /1MAA indexReleased
Sonar Deep Research128K$2$8n/a2025
Sonar Pro200K$3$15n/a2025
Sonar127K$1$1n/a2025

Privacy and data policy

  • On Hoodgen, prompts and attachments answer that request and stay in your conversation history so you can reopen it. They are not used to train models.
  • Requests route through OpenRouter to Perplexity and are covered by that provider's API terms.
  • Do not paste secrets or personal data you would not send to any hosted model.

Sonar Deep Research questions, answered

Sonar Deep Research is a reasoning model from Perplexity, released 7 March 2025. Sonar Deep Research is a research-focused model designed for multi-step retrieval, synthesis, and reasoning across complex topics.
✦

Start chatting with Sonar Deep Research

Free to try on Hoodgen AI, with a live credit estimate before every send. No API key, no setup.