Sonar Reasoning

About Sonar Reasoning

Sonar Reasoning is Perplexity's advanced model combining search augmentation with enhanced reasoning capability. It tackles complex queries that require both current information and multi-step analysis, synthesizing search results through sophisticated reasoning processes. The model excels at research questions requiring analysis of multiple sources, comparative assessments, and conclusions drawn from current data. Sonar Reasoning demonstrates the power of combining real-time information access with deliberate thinking, producing well-reasoned responses grounded in cited sources. It's particularly valuable for professional research, competitive analysis, and complex fact-checking tasks. The model represents Perplexity's advancement toward AI that doesn't just retrieve information but genuinely reasons about it. For applications requiring both accuracy and analytical depth, Sonar Reasoning offers unique capability at the intersection of search and reasoning AI.

Model Specifications

Perplexity Model Lineup

Compare all models from Perplexity to find the best fit

Similar Models from Other Providers

Cross-brand alternatives with similar capabilities

Gemma 3 4B

Google

Input: Free

Output: Free

Context: 96k

Qwen2.5-VL 7B Instruct

Alibaba Qwen

Input: Free

Output: Free

Context: 33k

Llama 3.1 405B Instruct

Meta

Input: Free

Output: Free

Context: 10k

Molmo2 8B (free)

Allen AI

Input: Free

Output: Free

Context: 37k

🚀 Quick Start

Get started with Sonar Reasoning API

from openai import OpenAI

client = OpenAI(
    base_url="https://api.provider.com/v1",
    api_key="YOUR_API_KEY"
)

response = client.chat.completions.create(
    model="perplexity/sonar-reasoning",
    messages=[
        {"role": "user", "content": "Hello!"}
    ]
)
print(response.choices[0].message.content)

Continue Your Decision