Sonar Deep Research, the Perplexity (Sonar) model in numbers
Sonar Deep Research in numbers: price per token, context window, independent benchmark scores and comparable models, based on public data checked on October 22, 2026.
Data as of October 22, 2026
Key figures
| Developer | Perplexity (Sonar) |
|---|---|
| Released | February 1, 2025 |
| Knowledge cutoff | January 2025 |
| Context window | 128,000 tokens (≈ 192 pages) |
| Maximum output | 32,768 tokens |
| Input price | $2 per million tokens (developer's price) |
| Output price | $8 per million tokens |
| Open weights | No |
| Reasoning mode | Yes |
| Tool calling | No |
| Structured output (JSON) | Not disclosed |
| Accepted inputs | text |
| Providers offering it | 5 |
What you need to know
Perplexity (Sonar) released Sonar Deep Research on February 1, 2025. It has a reasoning mode: it can "think" before answering, which helps with multi-step problems but uses more tokens, and therefore more money. Its weights are not published, so it only runs remotely, through the Perplexity (Sonar) API or partner hosts (5 listed on models.dev). That makes it more than a year old: successors may have shipped since.
We file it under mid-range models (34 models in this category on the site).
Pricing and budget
Sonar Deep Research costs $2 per million input tokens and $8 per million output tokens (developer's price, checked on October 22, 2026).
In practice, an agent task that reads 20,000 tokens and writes 2,000 costs about $0.056, or $56 for 1,000 such tasks.
That is 1.8 times the median output price of mid-range models ($4.38).
To estimate your own budget, use the agent cost calculator.
What do independent benchmarks say?
As of October 22, 2026, Epoch AI publishes no score for Sonar Deep Research on our basket's benchmarks. To place it, rely on its specifications and price, then test it on your own cases.
What it can do
Its context window reaches 128,000 tokens, roughly 192 pages of text at 500 words per page: enough for a medium-length document. A single answer can run to 32,768 tokens.
It accepts as input: text. Its knowledge stops at January 2025: beyond that, it needs web search or your documents.
It is not listed as able to call tools, so it is a poor fit for driving an agent that has to act on other software.
When to choose it
Given its category, Sonar Deep Research is mainly aimed at a balance between quality and cost, for everyday uses that do not justify a frontier model.
Before adopting it, test it on your own examples: benchmarks measure standard tasks, not your job.
What could replace it?
In the same category, the models closest to Sonar Deep Research are:
- Mistral Medium 3.5 (Mistral AI, $7.5 per million output tokens)
- Gemini 3.5 Flash (Google DeepMind, index 76)
- Claude Sonnet 5 (Anthropic, index 79.3)
- GPT-5.1 (OpenAI, index 75.8)
Similar models
- Mistral Medium 3.5 · Mistral AI
- Gemini 3.5 Flash · Google DeepMind · index 76
- Claude Sonnet 5 · Anthropic · index 79.3
- GPT-5.1 · OpenAI · index 75.8
Frequently asked questions
How much does Sonar Deep Research cost?
As of October 22, 2026, Sonar Deep Research costs $2 per million input tokens and $8 per million output tokens (developer's price). A task that reads 20,000 tokens and writes 2,000 costs about $0.056.
What is the context window of Sonar Deep Research?
128,000 tokens, roughly 192 pages of text. An answer can reach 32,768 tokens.
Is Sonar Deep Research open source?
No. It is a closed Perplexity (Sonar) model, available only through an API.
When was Sonar Deep Research released?
Sonar Deep Research was released on February 1, 2025, according to models.dev.
Can you build an agent with Sonar Deep Research?
Not recommended: it is not listed as able to call tools.
Data sources
- Models.dev — Model data: models.dev (open-source project)
- OpenRouter — Market prices cross-checked with OpenRouter
Data as of October 22, 2026
Report an error
Does a figure, a date or a fact look wrong? Tell us: every report is checked.