The Gemini 3 Pro family
Gemini 3.1 Pro does not carry one rate; it carries eight active rates: four service tiers, each split into two steps at the 200,000-token line. The cheapest configuration is $1 in and $6 out, and the costliest is $7.20 and $32.40, for the same token volume.
- four independent service tiers
- a 200,000-token step boundary
- unpublished context-window parameter
Last checked: This is a reference page. It is re-checked against the vendor sources and updated when a new version ships. Change log
Eight rates that all belong to one model
Google pricing page logs four independent tiers for this model. The Standard tier is $2 input and $12 output below 200,000 tokens, rising to $4 and $18 above that line. The Batch tier runs at exactly half the Standard rate: $1 and $6, moving to $2 and $9 past the line. The Flex tier carries the same rates as Batch. The Priority tier is $3.60 and $21.60, rising to $7.20 and $32.40 past the line.
The spread between the cheapest and costliest configuration, for a long request, is 3.6 times. None of these eight figures is incorrect, and none of them alone equals the rate for Gemini 3.1 Pro. If a single rate has been quoted for this model anywhere, that source did not specify which tier and which side of the boundary it logged.
The 200,000-token boundary is your most expensive architecture decision
Crossing the 200,000-token boundary doubles the input rate and multiplies the output rate by 1.5. So if you are building an architecture that resends an entire conversation history or an entire code repository on every call, the day your input crosses that boundary the invoice steps up with nothing in your code having changed.
The operational fix is to measure input length before sending and keep it under the boundary where feasible. This class of optimization is invisible in the unit rate and only surfaces on next month invoice.
The parameter Google does not publish
Google models page logs no figure for context window on gemini-3.1-pro-preview. All four Chinese vendors in this reference publish this parameter; Alibaba goes further and reports maximum input and chain-of-thought ceiling separately, and Google does not publish even the simplest form of this parameter.
The consequence for our table is direct: the context-window column for this candidate stays empty and scores nothing. We do not fill that cell with a plausible figure, because an unsourced fact in a comparison table is less reliable than an empty cell. Notably, this same pricing page announces the 200,000-token step boundary while never stating the overall window ceiling.
The family name and the model actually deployed
This family is named 3 Pro, but its only active text model is 3.1 Pro, still in preview status. No model with the identifier gemini-3-pro-preview exists on Google pricing page. The only entity still carrying the 3 Pro name is an image model, gemini-3-pro-image, whose status is stable and whose public name is Nano Banana Pro.
So if you arrived looking for Gemini 3 Pro, what is actually available to you is either 3.1 Pro in preview status or an image generation model. That image model inherits its text rate from 3.1 Pro, and its image output rate is $120 per million tokens, which computes to roughly $0.134 for a 1K or 2K image and $0.24 for a 4K image.
Google Search integration, and a shared quota
One further pricing row exists only for Google: integration with Google Search. The first 5,000 search requests per month are free; beyond that, each 1,000 requests costs $14. The detail that typically falls out of analysis is that this 5,000-request quota is shared across every model in the 3.x generation, not allocated per model. In architectures running several services across several Gemini models, all of them draw from a single shared quota.
Release timeline
Each version with its own release date, and what changed against the one before it.
- Gemini 3.1 Pro preview A tiered pricing structure and a preview-status identifier; Google has not published a context window specification for this version.
Using it from Iran
This column carries a date and says how it was checked, because most listicles guess it. Where we have only read a vendor policy page, the note below says exactly that.
- Reachable
- blocked
- Payment
- no working route
- Free tier
- no
- Measured on
How we checked: The Gemini API available-regions page logs the countries where the API is operational, and Iran is not on that list. This status is drawn from that document, not a direct network measurement.
Documented strengths
- The Batch tier runs at exactly half the Standard rate, so any workload without urgency halves in cost
- Native integration with Google Search, with a 5,000-request free monthly quota
- Four independent service tiers allow cost and priority to be tuned separately
- Context caching is priced on its own schedule, $0.20 below the 200,000-token boundary
Known weaknesses and limits
- Google publishes no figure for this model context window, so the context-window column in our table stays empty and scores nothing.
- The only active text model in this family remains in preview status, so neither its behaviour nor its rate carries a stability guarantee.
- The rate steps at the 200,000-token boundary and input doubles, so the cost of a long request is not linear.
- No free tier exists; the pricing page lists none for this model.
- The 5,000-request free search quota is shared across every model in the 3.x generation, not allocated to this model alone.
- The family is named 3 Pro while the active text model is 3.1 Pro, and the only entity still carrying the old name is an image model.
- Iran is not on the Gemini API available-regions list, so no official deployment route exists.
Technical verdict
Before comparing this model against any other, first determine which service tier you are deploying against, because the spread between the cheapest and costliest configuration is 3.6 times and most comparisons ignore this parameter. And if context window is your selection criterion, this model has no published figure for that parameter today.
Frequently raised questions, with documented answers
What does Gemini 3.1 Pro cost
It depends on the service tier and request length. The Standard tier is $2 and $12 below 200,000 tokens and $4 and $18 above it. The Batch tier is half the Standard rate, and the Priority tier is 1.8 times it. No single figure exists for this model.
What is the Gemini 3.1 Pro context window
Google logs no figure for this parameter on the Gemini API models page. We do not estimate it, so this cell is empty in our table. The same pricing page announces a 200,000-token step boundary, but that boundary applies to rate, not to the window ceiling.