Short answer
For a production-oriented Gemini 3 API choice, Google lists Gemini 3.8 Flash at $0.75 per 1 million input tokens and $3.75 per 1 million output tokens at its introductory price. Google’s model documentation identifies the stable API model ID as gemini-3.8-flash. Prices and terms can differ by platform: Google’s Gemini Enterprise Agent Platform price schedule says those rates apply through December 31, 2026, with standard rates of $1.50/$7.50 per million input/output tokens starting January 1, 2027. That dated schedule is for Agent Platform; don’t assume its change date applies to every API surface. (Google announcement; model documentation)
Example budget
For a hypothetical 1 million billable input tokens and 100,000 billable output tokens, the rate arithmetic is:
- Input: 1 × $0.75 = $0.75
- Output: 0.1 × $3.75 = $0.375
- Estimated total: $1.125
Google’s pricing documentation says processed tokens are billed, including thinking tokens and internal tool-system tokens. So these examples are rate arithmetic, not a complete workload forecast unless the quantities represent the API’s billable token counts. (Gemini Developer API pricing)
Which “Gemini 3” price do you mean?
| Model or channel | Price and status | What to know |
|---|---|---|
| Gemini 3.8 Flash — Gemini API | Google announced $0.75/1M input tokens and $3.75/1M output tokens at an introductory price. | The stable API model ID is gemini-3.8-flash. Check the live pricing page for the API surface and terms you plan to use. (Announcement; model documentation) |
| Gemini 3.8 Flash — Gemini Enterprise Agent Platform | $0.75/$3.75 per 1M input/output tokens through Dec. 31, 2026; $1.50/$7.50 from Jan. 1, 2027. Prices are in USD. | This dated price schedule applies to Google Cloud’s Agent Platform pricing. Do not automatically apply it to other Gemini API billing surfaces. (Google Cloud pricing) |
| Gemini 3 Flash Preview | Google’s published launch rates were $0.50/1M input tokens and $3/1M output tokens; audio input was listed separately at $1/1M tokens. | It is a preview model, gemini-3-flash-preview, not the same model as stable Gemini 3.8 Flash. Google says preview models may have more restrictive limits and may be deprecated; use the current pricing page before budgeting for it. (Google announcement; model documentation) |
| Gemini 3 Pro Preview | No current API price: the model was shut down on March 9, 2026. | Its original launch price—$2/1M input and $12/1M output tokens for prompts of 200,000 tokens or less—was historical, not a price for a currently callable model. (Model documentation) |
How to estimate your Gemini API spend
For a simple text-rate calculation, use:
Estimated token cost = (billable input tokens ÷ 1,000,000 × input rate) + (billable output tokens ÷ 1,000,000 × output rate)
For example, at Gemini 3.8 Flash’s announced introductory rates, 10 million hypothetical billable input tokens and 1 million hypothetical billable output tokens give $7.50 + $3.75 = $11.25. This is rate arithmetic—not a complete workload forecast unless those quantities match the API’s billable usage. Google says thinking tokens and internal tool-system tokens are billed too; don’t estimate from visible text alone. (Gemini Developer API pricing)
For a production budget, first identify the exact model ID and API surface, then estimate usage from representative requests and consult the API’s billed token counts. Recheck Google’s official pricing page before deployment, especially when budgeting across a promotional period or using a different Google Cloud product. Google describes the Gemini Developer API’s paid tier as intended for production applications and lists Batch API as a cost-reduction option; verify current model-specific eligibility and terms before relying on a discount. (Gemini Developer API pricing)
API billing is not a Gemini app subscription
The prices above are API usage rates, billed by token use—not monthly consumer-app subscription prices. Google presents Gemini API free and paid access separately; the paid API tier is intended for production applications. Don’t assume an app subscription includes API usage, or that API billing includes consumer subscription benefits. (Gemini Developer API pricing)
Bottom line: For a stable Gemini 3 production model, budget around Gemini 3.8 Flash’s stated per-token rates and confirm the live price for your chosen platform. Treat the older Gemini 3 Flash Preview rate as preview-model pricing, and don’t use Gemini 3 Pro Preview’s launch price for a model Google has shut down.
Sources
Introducing Gemini 3.8 Flash and 3.8 Flash Cyber
Gemini 3.8 Flash — Gemini API model documentation
Agent Platform Pricing — Google Cloud
Gemini 3 Flash Preview — Gemini API model documentation