Perplexity vs Gemini: both about $20, but Perplexity Pro lists no Gemini Pro model

Google AI Pro ($19.99 a month) wins on Gemini model access: its plan page lists Gemini 3.1 Pro, while Perplexity Pro ($20) lists only Gemini 3.7 Flash.

For an app, 1,000 answers cost $3.38 to $17.38 on Gemini 3.8 Flash and $4.25 on Perplexity Sonar, as of October 6, 2026.

Table 1. Plans compared by price, Gemini Pro model name, and usage limit as published. Rows run from highest to lowest price. Data collected 2026-10-06.

Table note: Bold marks the plan with a Gemini Pro model at about $20. Not named means the plan page names none. Perplexity model names follow the list dated 2026-09-04. The standard limit is the Gemini limit with no paid plan.

The Perplexity pricing page lists Pro at $20 a month and Max at $200 a month, and neither plan names a Gemini Pro model.

The Google AI plans page lists Plus at $4.99, Pro at $19.99, and Ultra at $99.99 or $199.99, and each plan lists a Gemini Pro model.

Google AI Pro costs $0.01 less than Perplexity Pro each month.

Where these numbers come from

Every price, limit, and model name here comes from a vendor page, collected on October 6, 2026 (UTC). LLM Waves Research collected these figures and did not run any of these models.

The cost figures are calculated from published rates, with every step shown below, and the full source list sits on the methodology page.

Google lists a Gemini Pro model on each paid plan, and Perplexity names none.

The count is zero for both Perplexity plans and one for each Google plan.

Perplexity Pro lists eight models, and Gemini 3.7 Flash is the only Gemini one

A model is the AI engine behind each answer, and a token is a small piece of a word. Gemini comes in a Pro line and a Flash line, and Flash costs less per token.

The Gemini pricing page for developers lists Gemini 3.8 Flash at $0.75 per million input tokens and Gemini 3.1 Pro Preview at $2.00 for prompts up to 200,000 tokens.

The Perplexity model list is a snapshot as of September 4, 2026. It names eight models for Pro: Sonar 2, GPT-5.6 Terra, Gemini 3.7 Flash, Claude Sonnet 5, Kimi K3, GLM 5.3, Grok 4.6 and Nemotron 3 Ultra.

Max adds GPT-5.6 Sol and Claude Opus 5.

Gemini 3.7 Flash is the only Gemini model on that list. An older Perplexity help page with no date still names Gemini 3.1 Pro under Pro, and the dated list is the one used here.

Perplexity Pro also names Claude Sonnet 5, and the Perplexity vs Claude comparison sets Perplexity against the Claude Pro plan.

Gemini limits refresh every 5 hours up to a weekly cap, and Perplexity Pro publishes no numbers

A usage limit is the amount of work a plan covers before you must wait. The Google usage limits page states that the Gemini limit refreshes every 5 hours until you reach a weekly limit, and that limits can change without notice.

AI Plus is 2x the standard limit, AI Pro is 4x, and AI Ultra is 5x or 20x the AI Pro limit.

Perplexity counts two kinds of work: Pro Searches and Research queries.

The plan comparison page lists Free at three Pro Searches a day and one Research query a month.

Pro gets weekly Pro Search limits and monthly Research limits, and Max gets the same kinds of limits for advanced use, with no numbers for either paid plan.

Only Google publishes a multiple for each paid plan.

Perplexity Max includes $100 of Computer credits a month, and Pro includes none

Computer is the Perplexity feature for tasks with many steps, and it runs on credits. Credits cost $1 for every 100, per the Perplexity credits page.

Pro has no monthly credits. It gets 4,000 credits once, and they expire in 30 days: 4,000 divided by 100 is $40. Max gets 10,000 credits each month, which is $100. Max also gets 35,000 credits once, which is $350, and they expire in 30 days too.

Only Max includes monthly credits, $100 of them, and Pro includes $0.

At 100 credits per $1, the Max credits equal half of the $200 price each month, and the Pro bonus equals twice the $20 price once.

Max also has a spending cap, set to $200 by default and adjustable up to $5,000 in the credit settings.

After the Pro bonus runs out, a Pro user buys credits in the same settings.

A batch of 1,000 answers costs $2.75 to $4.25 on Perplexity Sonar and $3.38 to $17.38 on Gemini 3.8 Flash

An API is the paid tool a developer uses to put a model inside an app. Input tokens are the words sent to the model, and output tokens are the words that come back.

The example here uses 2,000 input tokens and 500 output tokens per answer, with one web search per answer, so a different mix changes the total.

The two models priced are Perplexity Sonar and Gemini 3.8 Flash, each with web search.

Perplexity sells search at two prices, standard and fast, and the example shows both.

Perplexity Sonar stays between $2.75 and $4.25, and Gemini 3.8 Flash spans $3.38 to $17.38.

Perplexity Sonar

Perplexity Sonar (perplexity/sonar on the Agent API) costs $0.25 per million input tokens and $2.50 per million output tokens, per the Perplexity Agent API model list.

Its web search tool costs $2.50 per 1,000 calls, or $1.00 per 1,000 calls with the fast search option, per the web search tool guide.

• Input: 2,000,000 tokens at $0.25 per million is $0.50.

• Output: 500,000 tokens at $2.50 per million is $1.25.

• Search: 1,000 calls at $2.50 per 1,000 is $2.50, or $1.00 with fast search.

• Total: $4.25, or $2.75 with fast search.

The Perplexity pricing docs also list an older Sonar rate of $1 per million input and output tokens, and the example uses the Agent API rate.

The Perplexity Sonar model page on this site has other details on that model.

Gemini 3.8 Flash

Grounding means an answer draws on Google Search results. Gemini 3.8 Flash costs $0.75 per million input tokens and $3.75 per million output tokens through December 31, 2026.

The first 5,000 grounded searches a month are free, then $14 per 1,000, per the Gemini pricing page.

• Input: 2,000,000 tokens at $0.75 per million is $1.50.

• Output: 500,000 tokens at $3.75 per million is $1.875.

• Tokens together: $3.375, shown as $3.38.

• Search: 1,000 searches at $14 per 1,000 is $14.00, but only after the free 5,000.

• Total: $3.38 while free searches last, and $17.38 after.

Google bills every search the model runs, so an answer that runs two searches counts twice, per the Gemini search grounding guide.

Perplexity Sonar costs less past 5,333 answers a month, and at every volume from 2027

The break-even point is the monthly number of answers where two setups cost the same.

Both Gemini 3.8 Flash token rates double on January 1, 2027, a change the Gemini API pricing article also covers.

In 2026, a Gemini answer costs $0.003375 in tokens, and a Sonar answer costs $0.00425 in all.

Free searches cover the first 5,000 answers, so Gemini costs $4.375 less there: Sonar costs $21.25, and Gemini costs $16.875.

Each answer after 5,000 adds a $0.014 search, so Gemini costs $0.017375 per answer, which is $0.013125 more than Sonar.

Dividing $4.375 by $0.013125 gives 333, so the two lines cross at 5,333 answers a month.

From January 1, 2027, tokens alone cost $0.00675 per answer on Gemini 3.8 Flash, which is above the $0.00425 that Perplexity Sonar charges for a whole answer.

A batch of 1,000 answers costs $6.75 on Gemini with free searches and $20.75 without them.

The 2026 Gemini line crosses the Perplexity Sonar line between 5,000 and 6,000 answers, and the 2027 line stays above it.

Gemini trains on chats when Keep Activity is on, and Perplexity has data retention on by default

Training means a company uses saved chats to improve its models. Keep Activity is the Google setting that saves your Gemini chats.

With it on, Google uses them to improve its services, including training AI models, per the Gemini privacy page. With Keep Activity off, chats are saved for 72 hours and are not used for training unless you send feedback.

The Perplexity data page states that AI Data Retention is on by default for Free, Pro, and Max accounts. You switch it off under Account settings, then Preferences.

What we did not measure

Only prices, limits, model names, and data settings were collected. This article did not run a latency test, a throughput test, or an output-quality benchmark against any model discussed.

It does not rate accuracy, coding skill, or the quality of cited sources on either side. It leaves out Workspace and Enterprise plans, regional prices, taxes, annual prices, and student offers.

Which plan fits which reader

One plan at about $20

Top pick: Google AI Pro at $19.99 a month, for readers who want a Gemini Pro model, which the Perplexity model list does not name.

If you want models from several makers in one plan, this pick is wrong for you; use Perplexity Pro.

Situational: Perplexity Pro at $20 a month, only for readers who want Claude Sonnet 5, GPT-5.6 Terra, and six other models in one plan. If you need a Gemini Pro model, this pick is wrong for you; use Google AI Pro.

Heavy daily use

Situational: Google AI Ultra at $99.99 a month, only for readers whose main need is a higher Gemini limit, listed at 5x the AI Pro limit.

If you do not reach the AI Pro limit, this pick is wrong for you; use Google AI Pro at $19.99.

Insufficient data: Perplexity Max at $200 a month, because the plan comparison page publishes no limit numbers for it.

Developers and high volume

Best value: Perplexity Sonar with fast search at $2.75 per 1,000 answers, the lowest of the four setups, before any quality check.

If your answers must come from Google Search, this pick is wrong for you; use Gemini 3.8 Flash with grounding.

Situational: Gemini 3.8 Flash at $3.38 per 1,000 answers, only while its 5,000 free monthly searches last and the 2026 token rates apply.

If you run more than 5,333 answers a month, or any answers after January 1, 2027, this pick is wrong for you; use Perplexity Sonar.

Is Perplexity better than Gemini?

Neither plan is better on every measure collected. Google AI Pro lists a Gemini Pro model, Perplexity Pro names eight models from several makers, and Perplexity Sonar costs less than Gemini 3.8 Flash for 5,333 answers a month. Answer quality was not tested.

Does Perplexity use Gemini?

Yes. Perplexity Pro lists Gemini 3.7 Flash among its eight models, and Perplexity Max lists it among its ten. The same list names no Gemini Pro model.

Which is cheaper, Perplexity or Gemini?

Perplexity Pro costs $20 a month and Google AI Pro costs $19.99, a gap of $0.01. For 1,000 answers through an API, Perplexity Sonar costs $4.25, or $2.75 with fast search, and Gemini 3.8 Flash costs $3.38 until free searches run out.

Is Perplexity or Gemini better for research?

The pages publish limits, not research quality. Perplexity Free gets one Research query a month, and Pro gets monthly Research limits with no number. Google states that Deep Research, its Gemini feature for long reports, uses more of the Gemini limit.

Neither side was tested for accuracy.

What is the difference between Perplexity Pro and Gemini Pro?

Perplexity Pro is a plan that costs $20 a month. Gemini Pro is a line of Gemini models, not a plan. Google sells it through Google AI Pro at $19.99, so that is the plan to set against Perplexity Pro.

Plan pages and rates can change before you pay

Google states that limits can change without notice, and the Gemini 3.8 Flash price doubles on January 1, 2027.