Sonar Pro Search: What It Does and What It Costs
perplexity/sonar-pro-searchSonar Pro Search is Perplexity’s text-and-image model for deeper search reasoning, available through the OpenRouter API.
200K tokens of context — about 150,000 words, on a scale that ends at 2M.
What it can do
- Reasoning
- Vision
- Takes text
- Takes image
- Gives text
Released 30 Oct 2025 · checked 3 Oct 2026.
About it
Sonar Pro Search is a Perplexity model for deeper reasoning and analysis. It can take text or images and returns text. It has a 200,000-token context window and can write up to 8,000 tokens in one answer. It is available through the OpenRouter API, with separate prices for input and output.
Review
In a line
Sonar Pro Search is Perplexity’s model for deeper reasoning and analysis. It accepts text and images and responds with text.
What it's good at
The model is suited to search tasks that need deeper reasoning or analysis. It supports image input as well as text, and its reasoning capability is enabled.
Price in plain terms
Input costs $3 per million tokens, and output costs $15 per million tokens. Tokens are small units of text the model reads or writes; output costs more than input.
At the listed rates, reading a 500-word page costs about 0.19 cents. A chat turn with about 1,000 tokens in and 500 out costs about $0.01; 1,000 such turns cost about $10.50.
Context and limits
It can take up to 200,000 tokens of context at once—the material it considers for a response—and produce up to 8,000 tokens in one answer.
It returns text, not image output.
Ways to use it
Sonar Pro Search is available through the OpenRouter API. Its model ID is perplexity/sonar-pro-search. It takes text or image input and returns text.
Compare it with other Perplexity models, or browse reasoning models and vision language models.
Verdict
Consider it if you need a model for deeper search analysis, want to send images as well as text, and can use it through OpenRouter’s API. Its large context window may suit material-heavy tasks, though it can write no more than 8,000 tokens per answer.
It may not fit if you need image output or prefer access outside the OpenRouter API. Factor in the higher output rate when your use involves long answers or many turns.
Verdict
Sonar Pro Search is worth considering for people who need deeper reasoning for search and analysis, especially when their input may include images or a large amount of material. It is less suitable for anyone who needs the model to create images or wants access outside OpenRouter’s API. The input price is modest compared with output, so frequent or lengthy responses can account for much of the bill. Estimate usage with the per-turn example before choosing it.
- Accepts both text and image input.
- Supports deeper reasoning and analysis.
- Has a 200,000-token context window.
- Can produce up to 8,000 tokens in one answer.
- Available exclusively through the OpenRouter API.
- Returns text only.
- Output tokens cost more than input tokens.
- Pick it ifSearch and analysis that call for deeper reasoning; People who need to provide text or image input
- Skip it ifYou need image output.; You need access outside the OpenRouter API.
Pricing
Input is $3 per million tokens and output is $15 per million tokens. The example cost is about 0.19 cents to read a page, about $0.01 per chat turn, or about $10.50 for 1,000 such turns.
Questions
What is Sonar Pro Search?
It is a Perplexity model designed for deeper reasoning and analysis in search tasks.
Can it take images?
Yes. It accepts text and image input, but returns text.
Where can I use it?
It is available exclusively through the OpenRouter API. Its model ID is perplexity/sonar-pro-search.
How much can it read at once?
Its context window is 200,000 tokens.
How long can one answer be?
It can produce up to 8,000 tokens in one answer.
What does it cost?
Input costs $3 per million tokens and output $15 per million. A sample chat turn costs about $0.01; 1,000 such turns cost about $10.50.