How Much Does Gemini 3.8 Flash Cost?
Gemini 3.8 Flash is reported to run at $0.75 per million input tokens and $3.75 per million output tokens, the same rate as Gemini 3.6 and 3.7 Flash. That figure comes from third-party coverage rather than a dedicated Google rate card, so read it as reported, not confirmed.
The number Google does confirm sits underneath it. Gemini 3.6 Flash, released July 21, 2026, is listed at $0.75 input and $3.75 output as an introductory rate valid through December 31, 2026. After that the standard rate is $1.50 and $7.50. Whether the same expiry attaches to 3.7 and 3.8 Flash has not been published.
Budget next year off today's invoice and the confirmed half of the rate card already says the invoice is wrong by a factor of two. AI Perks tracks $7.7M in credits across 194 companies, the other lever on that number.
Every Gemini Flash Rate Google Has Actually Published
Two Flash-class rates are confirmed on Google's own pricing page. The rest of the line is either reported by third parties or not published at all.
| Model | Released | Input / 1M | Output / 1M | Status |
|---|---|---|---|---|
| Gemini 3.6 Flash | Jul 21, 2026 | $0.75 | $3.75 | Confirmed, introductory through Dec 31, 2026 |
| Gemini 3.5 Flash-Lite | Jul 21, 2026 | $0.30 | $2.50 | Confirmed, no expiry published |
| Gemini 3.7 Flash | reported Aug 13, 2026 | $0.75 | $3.75 | Reported as matching 3.6 Flash |
| Gemini 3.8 Flash | not published | $0.75 | $3.75 | Reported as matching 3.6 Flash |
| Gemini 3.5 Flash Cyber | around Jul 21 to 22, 2026 | not published | not published | Limited access, no public rate |
The only January 1 change Google has put in writing is the one on Gemini 3.6 Flash: $0.75 becomes $1.50 and $3.75 becomes $7.50. Gemini 3.5 Flash-Lite carries no expiry note at all, and no post-expiry rate has been published for 3.7 or 3.8 Flash. The three newest Flash models appear to sit at one price, but only one of those price tags has a Google page behind it.
What the January 1 Cliff Costs You
Double the per-token rate means double the bill, and the arithmetic is worth doing before December rather than in January.
Take 100 million input tokens and 20 million output tokens per month, a modest production agent, priced on the confirmed Gemini 3.6 Flash card:
| Scenario | Input cost | Output cost | Monthly total |
|---|---|---|---|
| Gemini 3.6 Flash, introductory rate today | $75.00 | $75.00 | $150.00 |
| Gemini 3.6 Flash, standard rate from Jan 1, 2027 | $150.00 | $150.00 | $300.00 |
| Gemini 3.5 Flash-Lite, no published expiry | $30.00 | $50.00 | $80.00 |
The third row is the useful one. Flash-Lite already costs roughly half of what Flash costs today, and because Google has published no expiry on the Flash-Lite rate, it is the one line of the bill with no scheduled increase behind it.
The cliff also widens the gap: on input, Flash is 2.5x Flash-Lite today and 5x once the window closes. Routing decisions made on today's ratio will be wrong in January.
What Actually Changed Across the Flash Line
Gemini 3.6 Flash, 3.7 Flash and 3.8 Flash all sit at the same reported rate. The differences between them are in efficiency and specialisation, not in the rate card.
Gemini 3.6 Flash arrived July 21, 2026 as the family's stated workhorse, with improved coding, knowledge-work and multimodal performance and, as TechCrunch reported, about 17% lower token usage than Gemini 3.5 Flash on comparable work. That is a real discount that never appears on a pricing page, because the saving is in tokens consumed rather than dollars per token.
Gemini 3.7 Flash is reported to have followed on August 13, 2026, positioned for everyday coding and agentic tool use at the same rate and the same context and output ceiling. Google has not published a release date or a separate rate card for 3.8 Flash.
The Cyber line is the one to be careful about. Gemini 3.5 Flash Cyber, introduced around July 21 to 22, 2026, has no published API price, because Google describes it as a limited-access model for governments and trusted partners rather than a general release. It is a cybersecurity model built on 3.5 Flash for finding and patching vulnerabilities, used inside Google's CodeMender agent. Any per-token price quoted for a Cyber model is a number that does not exist publicly.
The Flash Rates That Are Not Plain Text Tokens
Two Flash-branded models bill on something other than a text token, and both are easy to miss when you budget off the text rate card.
Gemini Omni Flash, announced at Google I/O on May 19, 2026, was the first shipping version of Google's any-to-any family: it accepts text, image, audio and video, and generates finished video. Google's pricing page lists $1.50 per million input tokens across text, image, video and audio, $9 per million text output tokens, and $17.50 per million video output tokens, roughly $0.10 per second at 720p. Video output is the line that breaks budgets, at nearly double the text output rate on the same model.
Omni Flash also reaches consumers without an API bill: it is bundled into the Gemini app across the Free, AI Plus ($4.99), Pro ($19.99) and Ultra ($100) plans, and is free through YouTube Shorts and the Create app. API pricing was not public at the initial announcement.
Nano Banana 2, listed as Gemini 3.1 Flash Image, is the image side of the same idea: Pro-level generation and editing at Flash speed. Google's pricing page lists $0.50 per million input tokens and $60 per million output tokens, which works out to roughly $0.045 to $0.151 per image depending on whether you render at 0.5K or 4K, with batch pricing at half rate. Its exact launch date is not confirmed, though a Nano Banana 2 Lite variant appears in Google's own blog listing under June 2026.
Those two rate cards are why a Gemini bill rarely matches a Flash text estimate. AI Perks tracks the credit programs that offset all of them.
What Google Has Not Published
Four things people keep asserting about this line are not on any Google page, and saying so plainly is more useful than filling the gap.
A dedicated 3.8 Flash rate card. The $0.75 and $3.75 figures for 3.8 Flash come from third-party coverage describing it as matching 3.6 and 3.7 Flash. Google has not published a standalone 3.8 Flash price or a release date.
Whether Flash-Lite is also on the clock. Gemini 3.5 Flash-Lite is listed at $0.30 and $2.50 with no expiry note attached. Google has published no date for a Flash-Lite increase.
Whether the introductory window will be extended. Google has published December 31, 2026 and nothing else. There is no announced extension and no announced grandfathering for existing contracts.
Cyber-tier pricing. Not published, as above.
The honest position on all four: the vendor has not published this. Model pricing moves fast enough that a page claiming certainty here is a page that will be wrong by November. AI Perks re-checks these rates rather than freezing a launch-day number.
How to Cut the Gemini Bill Before January
Three of these survive the price change, because they reduce tokens rather than rate.
Route by tier, not by habit. Flash-Lite at $0.30 and $2.50 handles classification, extraction and routing perfectly well, and carries no published expiry. Reserve Flash for work that needs Flash.
Upgrade for token efficiency, not for the headline. Moving from 3.5 Flash to 3.6 Flash was reported to cut token usage by about 17% on comparable work. At an identical rate, fewer tokens is a straight discount.
Watch what your multimodal calls bill. Omni Flash video output is listed at $17.50 per million tokens against $9 for text, and Nano Banana 2 output at $60 per million, with batch at half that. A pipeline that quietly generates media is not priced like one that generates text.
Offset the rest with cloud credits. Google runs a startup cloud program whose credits apply against Gemini API spend directly. No single universal amount is published and allocations are not uniform, so the figure that matters is the one your own profile draws.
Which programs stack, and in what order, is the part worth getting right, and it is what AI Perks exists to cover.
Frequently Asked Questions
Is Gemini 3.8 Flash free to use?
Google lists a free tier for Gemini 3.6 Flash, Gemini 3.7 Flash and Gemini 3.5 Flash-Lite. Specific limits are not reproduced here because Google revises them, and no separate free-tier listing has been published for 3.8 Flash. Production volume moves you to the paid tier either way, where credits from AI Perks offset the bill.
Does Gemini Flash pricing really double in 2027?
For Gemini 3.6 Flash, yes, per Google's own pricing page: input goes from $0.75 to $1.50 per million tokens and output from $3.75 to $7.50 once the introductory window closes on December 31, 2026. That is a discount expiring, not a surcharge. Google has not published whether the same expiry applies to 3.7 or 3.8 Flash.
What is the cheapest Gemini model right now?
Gemini 3.5 Flash-Lite, at $0.30 per million input tokens and $2.50 per million output, is the lowest-cost tier Google lists, built for high-volume, low-latency agentic work. Google has published no expiry date on those rates, so Flash-Lite may become relatively cheaper still in January.
Is Gemini 3.8 Flash cheaper than 3.6 or 3.7 Flash?
No. All three are reported at the same $0.75 and $3.75, and only the 3.6 Flash figure is confirmed on Google's own page. Choose between them on capability and token efficiency rather than price. Gemini 3.6 Flash was reported to use about 17% fewer tokens than 3.5 Flash.
How much does Gemini 3.5 Flash Cyber cost?
Google has not published a price. The Cyber variant is described as a limited-access model for governments and trusted partners rather than a general API release, so no public per-token rate exists. Any figure quoted for it is unverified.
Can startup credits cover Gemini API costs?
Yes. Google runs a startup cloud program whose credits apply against Gemini API spend. Amounts are not published as one universal number and vary by profile. AI Perks tracks $7.7M in credits across 194 companies, including the Google programs.
The confirmed rate doubles on January 1. Your credit balance does not have to.