Sonar
perplexity The API that serves this model and sets the prices shown here. sonar The model identifier to send in API requests. LLM Generates text from a prompt, so it can be compared with other LLMs on price and capability. Medium The balanced tier of a model family, trading some capability for speed and cost. Retired Sep 27, 2026 This model has been retired and can no longer be relied on.Provider: Perplexity
Knowledge cutoff: September 2025
Sonar will be supported until September 27, 2026. Official migration guide provided. Deprecation announced on Perplexity models documentation page.
Sticker prices & cost-per-token
List rates as published by Perplexity, before any volume or enterprise discounts.
$0.25
$2.50
128,000 tokens
4,096
tokens per response (includes reasoning tokens)
Closed weights
$0.0625
Get this data free
Want to stay updated on price changes on this model? Pull this data free (no API key required) here:
curl "https://coolhandlabs.com/api/v2/inference_apis?q%5Bmodel_eq%5D=sonar&q%5Bsource_api_eq%5D=perplexity"Pricing History
| Effective | Price |
|---|---|
| Oct 5, 2026 – present | Input cost: $0.25 / 1M, Output cost: $2.50 / 1M, Cached input cost: $0.0625 / 1M, Cache creation cost: $0.00 / 1M, 5-minute cache write cost: $0.00 / 1M, 1-hour cache write cost: $0.00 / 1M, Reasoning output cost: $0.00 / 1M, Cache storage cost: $0.00 / 1M/hr |
History
Last change detected October 5, 2026 (20 days ago) by an automated scan.
| Date | Changed by | What changed |
|---|---|---|
| Sep 19, 2026 | Unknown | Initial record created |
Source
https://docs.perplexity.ai/getting-started/pricingComparable models
No comparable models identified yet.
Reliability
No reliability incidents recorded in the last 90 days.
Frequently asked questions
Sonar (Perplexity) costs $0.25 per 1M input tokens and $2.50 per 1M output tokens, as last published by Perplexity.
Sonar has a context window of 128,000 tokens (the maximum combined input and output it can handle in a single request).
Sonar sits in the medium size tier. Size tiers run from micro (small enough to run on a device) through small, medium and large to extra large (frontier flagships), and are assigned from the model's family, naming and limits rather than its price.
Sonar can generate up to 4,096 tokens in a single response. Reasoning (thinking) tokens count toward this limit.
Sonar's training data runs through September 2025, so it may not know about later events. Coolhand records the cutoff at month precision, since providers often publish only the month.
Sonar accepts text as input and produces text as output, as recorded in the Coolhand catalog.
No — the weights of Sonar have not been released, so it is only available through hosted APIs.
Yes — Perplexity deprecated Sonar on September 27, 2026. Sonar will be supported until September 27, 2026. Official migration guide provided. Deprecation announced on Perplexity models documentation page.
Perplexity retired Sonar on September 27, 2026.
Call GET https://coolhandlabs.com/api/v2/inference_apis?q[source_api_eq]=perplexity&q[model_eq]=sonar — free, no API key required, and scoped to just this model since source_api and model together are unique. See https://coolhandlabs.com/docs for the full REST API reference.
Track LLM pricing and deprecations automatically.
Start free