What LLM Does Perplexity Use? Full 2026 Model Breakdown

Perplexity uses its own Sonar models alongside models from OpenAI, Anthropic, Google, xAI, Moonshot AI, Z.ai, and NVIDIA. Its published Search lineup includes Sonar 2, GPT-5.6 Terra, Claude Sonnet 5, Gemini 3.7 Flash, Grok 4.6, and Kimi K3; personal Max adds GPT-5.6 Sol and Claude Opus 5. The official Search model table is dated September 4, 2026; the separate API catalogue, accessed October 11, also lists newer releases such as GPT-6.1 Sol and Claude Sonnet 5.5.

The important buying question is which model you can actually select for your task. Compare the Search and API lists below, or use GlobalGPT’s model catalogue to explore a separate workspace for research, writing, and coding.

On this page: Search models by plan · Sonar explained · Best, Thinking, and Research · API models · Which to choose · FAQ

Which LLMs Can You Choose in Perplexity Search?

Perplexity’s advanced-model availability page separates the personal Pro and Max plans from Enterprise plans. The following is its September 4, 2026 Search snapshot, accessed October 11, 2026. These are published plan entitlements; trials, staged rollouts, regional requirements, and organization settings can change an individual account’s selector.

For everyday paid Search, Pro already covers a broad selection. The model-access reason to consider personal Max is specific: GPT-5.6 Sol and Claude Opus 5 are Max-only in this table. Compare the wider Perplexity subscription benefits if you also need higher research allowances or other premium features.

Enterprise Pro and Enterprise Max list all ten models above, including GPT-5.6 Sol and Claude Opus 5. Administrators can restrict providers or individual models, so a shorter company selector may reflect policy rather than a missing subscription. Our Perplexity Enterprise features guide covers the team controls behind that distinction.

What Is Sonar, and Is It Perplexity’s Only Model?

Sonar is Perplexity’s own search-oriented model offering. Sonar 2 is the name in the published consumer Search table, while the Agent API uses the identifier perplexity/sonar. Those labels belong to different product interfaces; neither proves that every Perplexity response is generated by the same model.

An answer involves more than text generation. Perplexity retrieves material, passes relevant information into a model, and presents an answer with sources. Selecting Claude or GPT changes the model available to that workflow; it does not turn Perplexity into the full Claude or ChatGPT application. That distinction explains how Perplexity can use OpenAI models without being ChatGPT.

For example, a question about a company’s latest earnings needs the correct filing and reporting period before it needs elegant prose. A model can produce a fluent answer from the wrong quarter’s document. Open the cited report and check the date and figures when the result informs a decision.

Best, Thinking, Deep Research, and Model Council: What Changes?

In Perplexity’s current model-selection explanation, Best is the automatic model option inside Search. Thinking is a reasoning setting on supported models. Deep Research and Model Council are different workflows with their own model-selection rules.

Ordinary Search should not be described as running every listed LLM simultaneously. Model Council is the explicit multi-model workflow. Likewise, a news question does not prove that Best selected Grok, and an uploaded image does not prove that it selected Gemini. The task’s subject alone cannot identify the model.

Start with the result you need: a brief answer, a carefully reasoned explanation, or a longer report. The examples in what Perplexity AI is good for can help you choose a workflow before choosing a model.

Which Models Does the Perplexity API Support?

As of October 11, 2026, the Perplexity Agent API catalogue lists models beyond the consumer Search snapshot above. Examples include GPT-6.1 Sol, Claude Opus 5.5, Claude Sonnet 5.5, Gemini 3.8 Flash, and Grok 4.7. An API listing does not add that model to a Pro or Max Search subscription.

The Agent API combines a selected model with tools such as web search and URL fetching. Perplexity’s Sonar migration guidance describes Agent API presets for tasks ranging from short lookups to extended agent work. A preset can bundle model and tool settings, so the preset name and the underlying model are separate pieces of information.

What Does Model Access Cost?

Consumer subscriptions pay for access inside Perplexity’s app and website. Developer requests are metered separately. Perplexity’s subscription comparison identifies the API as a pay-as-you-go product and states that it does not include complimentary API credits.

For developers, these are example default-tier token prices per one million tokens in the Agent API catalogue on October 11, 2026. They exclude tool calls, caching adjustments, and any other applicable charges.

A token price is not the complete cost of a research request. A longer answer, more source material, or repeated tool use can change the bill even when the model stays the same. The Perplexity API cost breakdown explains how to compare those components.

How to Choose a Model for Your Task

Choose a Search model from the plan you actually have, then compare it on a task you repeat. A useful comparison holds the question, source material, and requested format constant. Look for errors, missed requirements, and the amount of editing needed rather than choosing by brand name alone.

• Quick factual lookup: start with Best and inspect the supporting source. Selecting the most expensive model is unnecessary if the answer is a date or a short definition.
• Writing from research: try Claude Sonnet 5 or GPT-5.6 Terra in Search. Ask for a clear audience, word limit, and source-linked factual claims; compare which draft needs less revision.
• Complex analysis: enable Thinking where supported. Personal Max users can also compare GPT-5.6 Sol and Claude Opus 5 on the same problem.
• A detailed research report: choose Deep Research. Its automatic model selection differs from manually choosing a Search model.
• Several perspectives on one decision: use Model Council on an eligible web plan and inspect disagreements in the underlying answers.

If your work is mostly turning research into long drafts, the Perplexity vs Claude comparison is more relevant than a generic model ranking. A provider’s own application may have project and editing features that do not come with its model inside Perplexity.

For document and multimodal workflows, compare Gemini and Perplexity’s product features. Model access, accepted input types, and the surrounding app tools are separate purchase decisions.

How to Select a Model and Diagnose Missing Options

1. Sign in to the Perplexity account that owns the subscription and start a Search query.
2. Open the model selector in the search box. Leave Best selected for automatic choice or choose a named model available to your plan.
3. If the selected model offers Thinking, enable it when the task benefits from additional reasoning time.
4. Submit the query and inspect the response together with its cited sources. Record the chosen model if you are comparing outputs.
5. If an option is missing, compare your plan against the Search table. On an Enterprise account, check with the administrator about provider restrictions.

A manual Search choice does not control the models behind Deep Research, Comet, or Computer. These products have separate selection rules. On the Free plan, Best can choose a model for you, but you cannot manually select an advanced Search model.

The same underlying model can produce different results in different applications because search tools, prompts, document handling, and context differ. The Perplexity vs ChatGPT comparison helps separate model capability from the features of the service you are buying.

When GlobalGPT Makes More Sense

If you want to switch between research, drafting, and coding models in one workspace, GlobalGPT is another option. Its live catalogue includes Perplexity alongside GPT-6.1 Sol, Claude Opus 5.5, Claude Sonnet 5.5, Gemini 3.8 Flash, Grok 4.7, and Kimi K3. These are GlobalGPT offerings, separate from the Perplexity Search subscription table.

A practical starting task is to research a question, save the source-backed notes, and give the same notes to a second model for drafting or critique. Compare how much work remains before the result is usable. View GlobalGPT pricing to match that workflow to a plan; access does not transfer your Perplexity Pro or Max subscription into GlobalGPT.

Frequently Asked Questions

What LLM does Perplexity use in 2026?

Perplexity combines Sonar with third-party models. Its September 2026 Search table includes Sonar 2, GPT-5.6 Terra, Claude Sonnet 5, Gemini 3.7 Flash, Grok 4.6, Kimi K3, GLM 5.3, and Nemotron 3 Ultra. Personal Max also lists GPT-5.6 Sol and Claude Opus 5.

Does Perplexity use GPT-6.1 Sol?

Perplexity’s Agent API catalogue lists openai/gpt-6.1-sol as of October 11, 2026. The published consumer Search snapshot instead lists GPT-5.6 Terra and GPT-5.6 Sol. API availability alone does not establish access in a personal Pro or Max model selector.

Which Claude models are available in Perplexity?

The published Search table lists Claude Sonnet 5 for personal Pro and Max, with Claude Opus 5 on personal Max. Both appear in the Enterprise tables. Separately, the Agent API lists newer entries including Claude Sonnet 5.5 and Claude Opus 5.5.

Can I choose a model on Perplexity Free?

Free users can use Best and a limited number of Pro Searches, but cannot manually choose an advanced Search model. Manual selection is a paid-plan feature. Thinking, Deep Research, and Model Council also have their own availability rules.

Does Perplexity run several models on every question?

No. Best automatically selects a suitable model inside Search. Model Council is the separate feature that runs several models on the same question and combines their answers; Perplexity lists it for Max and Enterprise Max on the web.

Is Sonar 2 the same as the Sonar API model?

Sonar 2 is the label in Perplexity’s consumer Search table. The Agent API catalogue lists perplexity/sonar. They belong to the same provider, but the labels alone do not establish identical versions, settings, pricing, or product entitlements.

Does Perplexity Pro include API usage?

A consumer Pro subscription and developer API billing are separate. Perplexity’s plan comparison says the API is pay-as-you-go with no complimentary API credits. A developer bill can include input tokens, output tokens, and tool calls.

Can I compare Perplexity and other models on GlobalGPT?

Yes. GlobalGPT’s catalogue lists Perplexity and other models, including GPT-6.1 Sol, Claude Sonnet 5.5, and Grok 4.7. You use a GlobalGPT account and its plan allowances; this does not transfer a Perplexity Pro or Max subscription or all native app features.