IonBin
All pages
Search
Model · Perplexity

Perplexity Sonar and Agent API

Perplexity's Agent API gives developers one interface to models from several providers, plus Perplexity's own search-grounded Sonar model. Perplexity ended support for the older Sonar Chat Completions API on September 27, 2026, and points users to Agent API presets instead. Sonar is listed at $0.25 input and $2.50 output per million tokens.

Verified Oct 18, 2026 Re-verify: past review date 4 sources

What it is

The Agent API combines multiple providers’ models with optional research tools. Sonar is Perplexity’s search-grounded model within that interface. API billing and model configuration differ from the consumer app; do not assume identical results or included usage.

How it works

Choose a specific model or a preset such as fast, low, medium, high or xhigh. Configure tools and budget for both token usage and tool calls. Check the published capability and service-tier columns for your chosen model.

The migration guide maps legacy Sonar and Sonar Pro to fast, Sonar Reasoning Pro to low, and Sonar Deep Research to high. Test your response parsing, citations, streaming and background-job handling before replacing a production integration.

Benefits

A common API can combine model inference with documented research tools; evaluate whether this simplifies your application.

Limits and risks

  • Search can retrieve recent sources, but citations do not establish that every answer is accurate. Open the cited page and check whether it supports the specific claim.
  • Flex is best-effort capacity for eligible models. Setting it on Sonar does not establish a discount.
  • Retrying or falling back can change costs and output behavior. Record the served model, tier, token usage and tool invocations.
  • Model availability and configurations change. Pin supported explicit model IDs where appropriate and monitor deprecation notices.

This page reflects documentation, not an IonBin benchmark.

Checked specs

Snapshot from official sources on Oct 18, 2026. Model lineups and prices change often; this page is re-verified every 30 days.

Sonar API token price (USD, checked 2026-10-18)perplexity/sonar: $0.25 input, $2.50 output and $0.0625 cache read per million tokens [1]
Service tiersFlex 0.5× and Priority 2× token rates apply only to supporting models. Sonar lists default processing only. Unsupported tiers are ignored; inspect response.service_tier. [1]
Agent API tool fees (USD)Per 1,000 invocations: standard web search $2.50; Fast Search $1; fetch URL $0.50; people/finance search $5. Tokens cost extra. [2]
Legacy migrationSonar Chat Completions support ended September 27, 2026. Synchronous/streaming calls are gradually reformulated; asynchronous requests are unsupported. Move async work to Agent API background mode. [3]

Best-fit stacks

Stack 04

Marketing content & social

Research, draft, design and review content with an assistant and the publishing tools you already use.

Agent compatibility

API accessYesOfficial docs ↗
AppsPerplexity consumer app is a separate surface; API usage is billed through the API Console.
Tool in this entryHow agents connect
Perplexity MCPAPI/CLIFile import

Levels come from each tool's recorded, sourced compatibility. How the levels work.

Sources

  1. Agent API models · Perplexityofficial accessed Oct 18, 2026
  2. API pricing · Perplexityofficial accessed Oct 18, 2026
  3. Sonar migration guide · Perplexityofficial accessed Oct 18, 2026
  4. API FAQ · Perplexityofficial accessed Oct 18, 2026

Practitioner views are restated in our words and labeled as views, not measured results. Product facts come from official sources dated above.

Suggest an edit

IonBin's wiki isn't user-edited. Corrections go to our Directory Curator, who checks them against primary sources before anything changes. Spot something wrong or out of date on “Perplexity Sonar and Agent API”? Email us with a link to a primary source.

Suggest an edit by email or write to contact@ionbin.com