Perplexity API Pricing
Perplexity's API centers on web-grounded answers. On September 27, 2026 it ended its Sonar Chat Completions API, which billed tokens plus a per-request search fee. Its Agent API replaces it: each model bills its own token rates and every tool call - web search, URL fetch, sandbox - is metered separately. Perplexity's own model there is Sonar.
Sonar (Agent API)
Perplexity's own grounded search model on the Agent API, the one model Perplexity sells under its own name after the Sonar Chat Completions API ended on 2026-09-27.
Sonar LAUNCHED JUN 2026
The Sonar API listed Sonar at $1/M input and $1/M output until Perplexity ended it on 2026-09-27.
Sonar Pro LAUNCHED JUN 2026
The Sonar API listed Sonar Pro at $3/M input and $15/M output until Perplexity ended it on 2026-09-27.
Sonar Reasoning Pro LAUNCHED JUN 2026
The Sonar API listed Sonar Reasoning Pro at $2/M input and $8/M output until Perplexity ended it on 2026-09-27.
Sonar Deep Research LAUNCHED JUN 2026
The Sonar API listed Sonar Deep Research at $2/M input and $8/M output, plus $2/M citation tokens, $3/M reasoning tokens and $5 per 1K search queries, until Perplexity ended it on 2026-09-27.
| Model | Input /M | Output /M | Cached | Context | Max output | Vision | Tools | Tier |
|---|---|---|---|---|---|---|---|---|
| Sonar (Agent API) FLAGSHIP | $0.25 | $2.50 | $0.0625−75% | not listed | — | ✗ | ✓ | Active |
| Sonar RETIRED | — | — | — | Retired Sep 27, 2026 | — | ✗ | ✗ | Retired |
| Sonar Pro RETIRED | — | — | — | Retired Sep 27, 2026 | — | ✗ | ✗ | Retired |
| Sonar Reasoning Pro RETIRED | — | — | — | Retired Sep 27, 2026 | — | ✗ | ✗ | Retired |
| Sonar Deep Research RETIRED | — | — | — | Retired Sep 27, 2026 | — | ✗ | ✗ | Retired |
Pricing across the lineup.
Perplexity pricing notes.
AI//COST first added Perplexity after verifying the public pricing page on 2026-06-14.
perplexity/sonar) first read at $0.25/M input, $2.50/M output and $0.0625/M cache read — tools billed per callPerplexity direct API access for production workloads.
PRICINGCanonical source used by AI//COST for list prices and fee notes.
DOCSPer-model token rates, cache rates and service tiers for the Agent API live on the models page, not the pricing page.
Perplexity's Sonar API was built for web-grounded answers: its invoice combined token prices with request fees, Pro Search fees, search-query fees, citation tokens or reasoning tokens, depending on the model. Perplexity ended it on September 27, 2026 in favour of the Agent API, where each model bills its own token rates and tools are metered per call.
AI//COST keeps the four Sonar API rows as retired pages with their last listed rates, for invoice checks, and tracks Perplexity's own per-token model on the Agent API. Third-party models the Agent API resells are priced on their makers' pages.
OpenAI
Broad model catalog and API ecosystem; useful reference for general-purpose chat pricing.
SAFETY LABAnthropic
Claude is the major enterprise comparator for agentic and document-heavy workloads.
OPEN-WEIGHTAlibaba (Qwen)
Qwen provides low-cost open and proprietary options for text and code workloads.
GLM ARCHITECTZhipu (Z.ai / GLM)
The Beijing lab behind the GLM models, also operating internationally as Z.ai. Spun out of Tsinghua University in 2019, Zhipu ships strong coding/agentic models, multiple free tiers, and open weights — and was the first of China's "Six Tigers" to pursue an IPO. It also sits on the US Entity List.
MEDICAL-AI SPECIALISTBaichuan
The Beijing lab behind the Baichuan general models and the medical M-series. Founded in 2023 by Wang Xiaochuan (former CEO of search engine Sogou), Baichuan has pivoted hard toward healthcare AI — its open-source Baichuan-M3 tops several medical benchmarks — while keeping a general-purpose 4-series lineup.
CHINA CONSUMER LEADERByteDance (Doubao)
The Doubao model family from ByteDance — parent of TikTok and Douyin. Served through the Volcano Ark (火山方舟) platform on ByteDance's Volcano Engine cloud, Doubao powers China's most-used consumer AI app and undercuts most frontier labs on price.
Frequently asked.
Practical notes on Perplexity pricing and what AI//COST includes.
Q · 01 What happened to the Sonar API? +
fast preset in place of Sonar and Sonar Pro, low for Sonar Reasoning Pro and high for Sonar Deep Research. The four Sonar pages on this hub keep their last listed rates for reading old invoices.Q · 02 What does Perplexity's own model cost now? +
perplexity/sonar) lists at $0.25/M input, $2.50/M output and $0.0625/M cache read. Tools bill per call on top: web_search is $0.0025 per invocation, or $0.001 with Fast Search.Q · 03 What Perplexity prices are tracked here? +
Q · 04 Are custom enterprise prices included? +
Q · 05 Are tool and search fees included? +
web_search $0.0025 per invocation ($0.001 with Fast Search), fetch_url $0.0005, people_search and finance_search $0.005, and the code sandbox $0.03 per session. Presets call tools on your behalf, so their cost depends on how many searches a run makes.