Guides / gemini 3.1 pro preview pricing

Gemini 3.1 Pro Preview pricing, and the date nobody wrote down

Gemini 3.1 Pro Preview pricing is $2.00 per million input tokens and $12.00 per million output tokens. Preview rates carry no commitment, so a forecast built on this one ages faster than any other number in your model. Culpa, a local-first LLM cost, margin, and forecast ledger, prices every call from a dated price book, so a rate change lands as a diff rather than as a surprise.

Why this happens

Every other rate in this book is a published price with no stated end. A preview rate is a different kind of number: it exists to get the model tried, and it changes when the model reaches general availability. Teams build a forecast on it anyway, because it looks exactly like a normal rate on the page. There's precedent for what happens next. Claude Sonnet 5's introductory pricing rose 50% on a published date, and the workloads pinned to the old rate had that increase land with no action on their side.

What this usually looks like

A preview model is carrying production traffic and the forecast treats its rate as fixed.
Nobody has written down what happens to the budget if the rate rises at general availability.
The migration path off the preview model has never been costed or tested.
Your cost model stores one rate per model with no effective date beside it.

Free, no card, no account

Run the free Cost Leak Scan

It shows your most expensive conversation before you install anything.

Run a free scan

Mistakes that cost the most

Mistake	Why it hurts	Do instead
Forecasting a preview rate as though it were permanent.	The rate exists to encourage adoption, so the direction of travel is usually upward.	Forecast a band, and put the upper end at what the stable tier in the same family charges.
Putting production traffic on a preview model with no exit.	A rate change or a deprecation both land as an emergency rather than as a scheduled task.	Keep a tested fallback and know what the same workload costs on it.
Storing one rate per model with no effective date.	A dated change can't be modelled at all if the data structure has nowhere to put the date.	Store rates with effective dates and price each call at the rate in force when it was made.

Run this check tonight

List every preview or experimental model carrying production traffic.
Price this month's volume at the current preview rate and at the stable tier's rate.
Write down the gap. That's the exposure nobody has approved.
Confirm a tested fallback exists and that somebody has priced the same workload on it.
Check whether your cost model can even represent a rate that changes on a date.

The band around a preview forecast

Illustrative example

A workload of 200 million input and 40 million output tokens a month on Gemini 3.1 Pro Preview at real rates, with two plausible landing points at general availability. Volumes are modelled and the landing points are scenarios, not announcements.

At the preview rate: (200 x $2.00) + (40 x $12.00) = $400 + $480 = $880

If it settles at Gemini 2.5 Pro's rates of $1.25 and $10.00: $250 + $400 = $650, which is 26% cheaper

If it rises 50%, matching the precedent set by Claude Sonnet 5: $600 + $720 = $1,320

The forecast band is $650 to $1,320 around a current figure of $880

Nothing on the rate card narrows that band, because a preview rate carries no commitment

The honest forecast is a band twice as wide at the top as at the bottom, not a number. Publishing $880 as a plan figure states a precision the underlying rate has never offered.

Every number, with its confidence and source

Figure	What it means	Confidence	Source
$2.00 / $12.00 per million	Gemini 3.1 Pro Preview input and output rates, with no commitment to stay there	calculated	Culpa price book row for gemini/gemini-3.1-pro-preview, $0.002 and $0.012 per 1k, effective 2026-07-02, verified against Gemini API pricing. Shown here per million.
$650 to $1,320	modelled monthly cost band for the same workload if the preview rate settles lower or rises	estimated	Both endpoints from the teardown arithmetic. The lower uses real Gemini 2.5 Pro rates and the upper applies the 50% rise Claude Sonnet 5's introductory pricing took. Volumes are modelled and neither landing point is announced.

What a generic answer can’t know

A provider page shows today's rate and says nothing about your exposure to it changing. That depends on your volume, your mix and whether a tested fallback exists. Culpa prices your real volumes against a dated price book on your infrastructure, keeps your prompts and responses there, and counts the calls to run your plan.

Questions founders ask next

How much does Gemini 3.1 Pro Preview cost?

$2.00 per million input tokens and $12.00 per million output, with context caching at $0.20 per million. Rates verified against Google's pricing page, and preview rates carry no commitment to stay there.

Should I run production traffic on a preview model?

Only with a tested fallback and a forecast expressed as a band rather than a number. The rate exists to encourage adoption, so plan for it moving, and know in advance what the same workload costs on the stable tier.

Does a preview rate usually go up?

Neither direction is settled, and both have precedent. Claude Sonnet 5's introductory pricing rose 50% on a published date, while a preview model settling near an existing stable tier can land cheaper. Forecast the band, not the point.

On your infrastructure

Culpa runs on your infrastructure. Your prompts and responses never leave it. Culpa counts calls to run your plan, and it fails open, so if it ever breaks your app keeps running.

How Culpa works

Find the culprit. Not just the total.

Your dashboard shows what you spent. It stops short of who spent it. Culpa shows the conversation, the user and the feature behind it.

Your prompts stay local.

Culpa runs on your own infrastructure. What you send to a model reaches us at no point.

Every dollar has a name.

Follow any charge to the conversation, the user, the feature and the customer behind it.

See the bill before it lands.

Cost your next feature before you ship it. You get the likely bill and the worst case, at best, median, p90 and p99.

Three steps to your first answer.

Change one base URL.

Or drop in the Python or TypeScript library.

Find your most expensive conversation.

In the first session, not the first week.

Cost your next feature before you ship it.

Base URLhttp://localhost:4545/v1Your traffic keeps flowing if Culpa ever stops.

Why the bill went up

Example dashboard

Calls traced

418,209

across 3 projects

Spend this week

$378.41

+ $182 vs last week

Failed calls

312

74% retried, and you paid for all of them

+ $182 this week traced to one culprit

Spend over 14 days

user_384report_generatorconv_91fprompt_v1894,220 tokens3 retries$6.81

Most expensive users

user_384$38.42

user_119$21.07

user_562$14.90

user_204$8.30

user_871$5.10

Next week forecast

Best$180

Median$240

p90$310

p99$395

Graded against reality. Accuracy shown as results land.

Run the free Cost Leak Scan Start 14-day trial

Free, no card, no account

Run the free Cost Leak Scan

It shows your most expensive conversation before you install anything.

Run a free scan

Keep reading

Sources: Gemini API pricing. Last reviewed 2026-08-01, rates effective 2026-07-02. Plain text version.