Gemini API Pricing (November 2026): Free Tier, Paid Tier and Billing
APIsQuick answer: The Gemini Developer API has a Free tier and a Paid tier, plus an Enterprise option. New projects start free, with limited access to certain models. The Paid Tier adds higher rate limits, the most advanced models, context caching, a Batch API discount, and the rule that content is not used to improve Google’s products. Google bills per million tokens (and per tool, for some features). Today’s numbers are on the official Gemini API pricing page (checked November 18, 2026). This is not the same bill as Google AI Pro vs Ultra.
How do Free, Paid and Enterprise compare?
From Google’s pricing overview:
| Tier | Who it fits | What you get | Official page |
|---|---|---|---|
| Free | Developers and small projects getting started | Limited access to certain models; free input and output tokens up to free rate limits; Google AI Studio access; content is used to improve Google’s products | Pricing |
| Paid | Production apps that need volume and advanced features | Higher rate limits; context caching; Batch API at half the standard rate; access to Google’s most advanced models; content not used to improve Google’s products | Pricing |
| Enterprise | Larger deployments on the Gemini Enterprise Agent Platform | Everything in Paid, plus options such as dedicated support, advanced security and compliance, provisioned throughput and volume discounts | Pricing |
Google also notes that Google AI Studio usage itself is free of charge in available regions. The API still follows the Free or Paid tier attached to the Cloud project behind your key.
Which Gemini model tier should you use?
Google lists many models on the pricing page. A practical reading of that page, without reprinting rates:
| Model family (examples) | Best for | Free Tier | Notes from the official page |
|---|---|---|---|
Gemini 4 Carbon (gemini-4-carbon) |
Hardest reasoning, coding, long-decode agents | Not available | Every token is billed, including thinking tokens and internal tool instructions when you use grounding. Paid rates on this model change on December 1, 2026. |
Gemini 4 Argon (gemini-4-argon) |
Frontier reasoning with faster, cheaper turns than Carbon | Not available | Same Free Tier rule and the same December 1, 2026 rate change as Carbon. |
| Gemini 4 Hydrogen | Mid-tier Gemini 4 when you still want a Free Tier option | Input and output listed as free of charge | Implicit context caching only; explicit caching is not supported on the Gemini 4 models above. |
| Gemini 3.8 Flash | High-volume or latency-sensitive work; long-horizon agents | Input and output listed as free of charge | Paid promotional rates run through December 31, 2026, then change on January 1, 2027. |
| Image models (Nano Banana family) | Image generation and editing in the API | Varies by model; check the table | Consumer how-to: How to use Nano Banana. |
| Older Gemini 2.5 models | Existing integrations | Check the live table | Still listed; compare them against current Flash and Gemini 4 prices before you standardize on them. |
Two billing details that are easy to miss, both from the same pricing page:
- Thinking tokens count. On Gemini 4 Carbon and Argon, output pricing includes thinking tokens, and tool system instructions can add to the bill.
- Grounding is metered separately. Grounding with Google Search or Maps has a monthly free allowance on paid Gemini 3 and newer models, then a per-query charge, plus a daily cap on the Developer API. The live numbers are on the pricing page.
How does Gemini API billing work?
The billing guide says new accounts begin on the Free Tier. To reach the Paid Tier you link a billing account and complete Google’s prepaid or postpay setup (a minimum prepaid credit applies). You then move through Tier 1, Tier 2 and Tier 3 based on cumulative spend and account age.
Tiers, rate limits and billing caps are set at the billing account level, not per API key. Each paid tier has a billing cap, listed on that billing page.
In AI Studio, click Set up billing next to the project you want to upgrade. Google will ask you to create or choose a Cloud billing account, then finish prepay or postpay. If you abandon a prepay setup mid-flow, the billing docs warn that other projects on that billing account can lose API access until you complete setup or contact Cloud Billing Support.
Step-by-step key creation, including the 2026 auth-key rules, is in How to get a Gemini API key.
Is this the same as Google AI Pro or Ultra?
No. Google AI Pro and Ultra are consumer subscriptions for the Gemini app and related consumer products. The Gemini Developer API is billed on the Cloud project behind your API key, at the rates on the Developer API pricing page.
A Pro or Ultra subscription does not automatically cover API token usage, and paying for the API does not give you Ultra’s consumer perks. If you are choosing a Gemini app plan, start with that comparison. If you are shipping an app or script that calls the API, this page and the official pricing table are the ones that matter.
How can you cut Gemini API costs?
- Stay on the Free Tier while you prototype, if the models you need are listed as free of charge and you can accept the data-use rule and lower rate limits.
- Use Flash or Hydrogen for everyday volume; reserve Carbon or Argon for tasks that need the frontier models.
- Turn on the Batch API for jobs that can wait. Google lists it as a 50% cost reduction on the Paid Tier.
- Use context caching on the Paid Tier when you resend the same large prompt context.
- Watch grounding. Search and Maps queries have their own free allowances and per-query charges, on top of model tokens.
- Set spend caps and alerts in Cloud Billing, as described in the billing guide.
For a side-by-side of how other labs structure API prices, see OpenAI API pricing and Claude API pricing.
FAQ
Is the Gemini API free?
The key is free, and new projects start on the Free Tier, which gives limited access to certain models with free input and output tokens, up to free rate limits. Google says Free Tier content is used to improve its products. The most advanced models, higher limits, context caching and the Batch API discount need the Paid Tier.
Is Gemini API pricing the same as Google AI Pro or Ultra?
No. Google AI Pro and Ultra are consumer subscriptions for the Gemini app and related products. The Gemini Developer API is billed on the Google Cloud project behind your API key, per the official pricing and billing pages. Paying for one does not replace the other.
Which Gemini models are not on the Free Tier?
Google's pricing page says Gemini 4 Carbon and Gemini 4 Argon are not available on the Free Tier. Other listed models, including Gemini 4 Hydrogen and Gemini 3.8 Flash, show free-of-charge input and output on the Free Tier. Always check the live table, because availability changes.
How do I move from the Free Tier to Paid?
In Google AI Studio, click Set up billing next to the project, link a billing account, and complete the prepaid or postpay flow Google shows you. Paid tiers (Tier 1 to Tier 3) then unlock from cumulative spend and account age. Tiers, rate limits and billing caps are set at the billing account level.
Does the Batch API cost less?
Yes, on the Paid Tier. Google's pricing page lists Batch API as a 50% cost reduction versus standard rates. Context caching is also a Paid Tier feature. Current rates are on ai.google.dev/gemini-api/docs/pricing.
Sources
- Google AI for Developers: Gemini Developer API pricing (accessed 2026-11-18)
- Google AI for Developers: Gemini API billing (accessed 2026-11-18)
- Google AI for Developers: Gemini API keys (accessed 2026-11-18)