$ AI PriceCompare
← Back to API models
§ Perplexity · API model

Perplexity Sonar Pro API pricing

Sonar Pro is Perplexity's higher-end web-grounded model, billed on tokens plus search fees. Perplexity ended Sonar Chat Completions support on September 27, 2026 and now runs Sonar traffic through its Agent API, which changes how search is billed.

Updated November 2, 2026
§ Short answer

The short answer

The table below shows the Sonar Pro token rates this site tracks, but the bill also includes search fees, so token prices alone understate the cost. Since September 27, 2026, Perplexity serves Sonar requests through its Agent API, where web search is a separately priced tool; check Perplexity's model list before budgeting, because Sonar Pro isn't currently listed there as its own Agent API model.

§ API pricing

Rates and specs

Sonar Pro
ProviderPerplexity
TierSpecialist
Input / 1M tokens—
Output / 1M tokens—
Cache hit / 1M—
Cache write / 1M—
Context window200K
NotesNot separately priced on the Agent API: Sonar Chat Completions support ended Sep 27, 2026, Sonar Pro isn't on the Agent API model list, and Perplexity maps it to the Agent API "fast" effort level (billed at that setup's model token rates plus tool fees)

Perplexity official docs ↗

§ Family

Other Perplexity models

ModelTierInput / 1MOutput / 1MCache hit / 1MContextNotes
Sonar ProSpecialist———200KNot separately priced on the Agent API: Sonar Chat Completions support ended Sep 27, 2026, Sonar Pro isn't on the Agent API model list, and Perplexity maps it to the Agent API "fast" effort level (billed at that setup's model token rates plus tool fees)
SonarSpecialist$0.25$2.50$0.0625128KCheap web-grounded answers; now served via the Agent API (perplexity/sonar) after Sonar Chat Completions support ended Sep 27, 2026; billed per token plus per tool call (web_search $0.0025 per invocation, $0.001 with Fast Search)
§ Rivals

Cheapest specialist rivals

ModelTierInput / 1MOutput / 1MCache hit / 1MContextNotes
CodestralSpecialist$0.30$0.90$0.03128KCode generation
Grok Build 0.1Specialist$1.00$2.00$0.20256KCode tasks; replaced Grok Code Fast 1; ≥200K tokens: $2/$4, cache $0.40

What changed on September 27, 2026

Perplexity's migration guide says Sonar Chat Completions support ended on September 27, 2026. Synchronous and streaming Sonar requests keep working because they're being reformulated as Agent API requests, rolling out model by model. Asynchronous Sonar requests aren't reformulated and no longer work; move them to the Agent API's background mode. For a native migration, Perplexity suggests the fast effort level as the starting point for both Sonar and Sonar Pro, and fast turns on web search by default.

How search is billed now

On the Agent API you pay the model's per-token rates plus tool fees: web_search is charged per 1,000 invocations (with a cheaper Fast Search option), and fetch_url per call. The old search_context_size setting has no named equivalent; instead you cap how much search content comes back with max_tokens and max_tokens_per_page on the web_search tool. Those caps are now your main cost lever, because retrieved pages count as input tokens.

Budgeting before you migrate

Perplexity's Agent API model list currently shows perplexity/sonar but no Sonar Pro entry, so confirm which model and rates your reformulated Sonar Pro requests are billed at by checking a few responses' usage before scaling up. Rate limits come from usage tiers (Tier 0 to Tier 5) set by your lifetime credit purchases, not your current balance, and you keep a tier once reached. Enterprise buyers can also purchase through AWS Marketplace.

§ FAQ

Frequently asked questions

Is the Sonar Pro API still available?

Sonar Chat Completions support ended on September 27, 2026. Synchronous and streaming requests are being reformulated as Agent API requests; asynchronous Sonar requests no longer work.

Does Sonar Pro charge per request as well as per token?

Search is billed on top of tokens. On the Agent API, web_search is priced per 1,000 invocations and fetch_url per call, in addition to the model's token rates.

What should I use instead of Sonar Pro on the Agent API?

Perplexity's migration guide suggests the fast effort level as a starting point for Sonar Pro workloads, with web search enabled by default. You choose the model separately.

§ Keep reading

Related guides.

How we choose: prices come from each vendor's official pricing page and are re-checked every week. We take no affiliate fees, so rankings aren't for sale.