Sonar Reasoning Pro

Deprecated · Language, Reasoning

Sonar Reasoning Pro is a reasoning model in Perplexity's Sonar API, first listed in the API docs in early February 2025, that returns its chain of thought in the response. Perplexity's docs say it is powered by DeepSeek R1 and aimed at complex, multi-step topics; Perplexity recommended it in place of R1 1776 and Sonar Reasoning when those left the API. [1] [2] [3] [4] Primary source

Timeline of Sonar Reasoning Pro →

Claims and evidence

  • Released No dated announcement found. The API model list does not show sonar-reasoning-pro in the capture of 2025-01-30 and does in the next capture, of 2025-02-05, so a release on 2025-01-30 or 2025-01-31 cannot be ruled out. Secondary sources giving 2025-03-07 conflict with these captures. Primary source[1]Supported Models table [5]Supported Models table
  • Deprecated Changelog entry tagged Deprecation, labelled July 2026 (month only). The migration guide gives 2026-09-27 as the end of Sonar Chat Completions support. Primary source[4]July 2026: Sonar Chat Completions migration
  • Status DeprecatedSonar Chat Completions support ended on 2026-09-27. Synchronous and streaming requests keep working but are being reformulated as Agent API requests, rolling out gradually by model; the migration guide maps this model to the low preset. The Agent API Models page lists no id for this model. Primary source[6]introduction; How Sonar Chat Completions maps to Agent API presets [4]July 2026: Sonar Chat Completions migration
  • Derived from (other) DeepSeek-R1 (DeepSeek)Perplexity says it is powered by DeepSeek R1; the method is not stated. Primary source[7]
  • Change · Other Perplexity's docs describe it as powered by DeepSeek R1 with chain of thought.Compared with Sonar Reasoning Primary source[2]Model descriptions: sonar-reasoning-pro
  • Change · Other Has an 8k maximum output, a limit the docs state for this model and Sonar Pro but not for Sonar Reasoning.Compared with Sonar Reasoning Primary source[1]Supported Models table
  • Change · Other Gives on average twice as many citations as Sonar Reasoning, according to Perplexity's model page.Compared with Sonar Reasoning Archived primaryOriginal source unavailable[3]
  • Feature Reasoning modeOutputs a think section with reasoning tokens before the answer. Primary source[3]Model Type; Key Features [1]notes under the table [7]Reasoning Models
  • Feature RetrievalWeb search with citations; about twice as many citations on average as Sonar Reasoning. Archived primaryOriginal source unavailable[3]Key Features
  • Context window 127k tokensListed as 127k at launch with a maximum output of 8k; the model list from 2025-02-22 and the model page give 128k. Primary source[1]Supported Models table
  • Parameters Not disclosedNot disclosed by Perplexity. No claim madeNo source
  • Access API Archived primaryOriginal source unavailable[3]Test the Model

Lineage

Predecessors

No known predecessor.

Successors

No known successor.

Based on

  • DeepSeek-R1 · DeepSeek · 20 January 2025 · derived (other)

Variants and derived

None recorded.

Siblings

None recorded.

All ancestors

All descendants

None.

Variants

No variants recorded in this record.

Related AI Radar coverage

AI Radar coverage starts in June 2026; no coverage linked yet.

All model releases from Perplexity on AI Radar →