Text Generation ModelDeprecated

Sonar Small Online

Sonar Small Online is a web-connected text generation model from Perplexity built on Llama 3.1 with a 127,072-token context window.

Start Building with Sonar Small Online View All Models

Publisher

Perplexity

TypeText

Context Window127,072 tokens

Replaced bySonar

Try Sonar Small Online →

About Sonar Small Online

Web-connected text generation with citations

Sonar Small Online is a text generation model developed by Perplexity AI, built on Meta's Llama 3.1 architecture. Its full model identifier is llama-3.1-sonar-small-128k-online, and it was added to MindStudio in May 2024. The model is part of Perplexity's Sonar family, which is designed to combine large language model capabilities with real-time web access, allowing it to retrieve and cite current information during inference.

The model supports a context window of up to 127,072 tokens and a maximum response size of 28,000 tokens, making it suitable for tasks that require processing longer documents or multi-turn conversations. It includes configurable options to return citations and images alongside responses, which is useful for research assistance, fact-checking, and question answering grounded in live web sources. Note that this model has been marked as deprecated, so developers building new applications should check Perplexity's current model offerings for an actively maintained alternative.

Capabilities

What Sonar Small Online supports

Real-Time Web Search

Retrieves live information from the web during inference, enabling responses grounded in current sources rather than a static training snapshot.

Citation Return

Optionally returns source citations alongside generated text, configurable via the Return Citations input on MindStudio.

Image Return

Can return relevant images alongside text responses when the Return Images option is enabled, surfacing visual results from web retrieval.

Long Context Window

Supports up to 127,072 tokens of context, allowing lengthy documents or extended conversation histories to be processed in a single request.

Text Generation

Generates natural language responses for tasks such as summarization, question answering, and research assistance, with a maximum response size of 28,000 tokens.

Ready to build with Sonar Small Online?

Get Started Free

FAQ

Common questions about Sonar Small Online

What is the context window size for Sonar Small Online?

Sonar Small Online supports a context window of 127,072 tokens, with a maximum response size of 28,000 tokens.

Is Sonar Small Online still actively supported?

No. The model's status is listed as deprecated. Developers starting new projects should review Perplexity's current model card documentation for actively maintained alternatives.

Does Sonar Small Online have access to real-time information?

Yes. The model is an online variant, meaning it can retrieve information from the web during inference rather than relying solely on its training data.

Can Sonar Small Online return citations with its responses?

Yes. The model includes a configurable Return Citations option that, when enabled, includes source references alongside the generated text.

What is the pricing for Sonar Small Online?

Pricing information is not published in the available metadata. Check Perplexity's official documentation or MindStudio's pricing page for current cost details.

Resources