Gemini API free tier and pricing: what is free and what paid adds
Gemini API in 2026: which models are free, paid prices per million tokens, context caching, Batch, usage tiers and how to enable billing in AI Studio.
You can call Gemini for free today, with no credit card, and the only real cost is privacy on your prompts. The free tier includes the Flash and Flash-Lite models plus embeddings and text-to-speech, with rate limits set per project and Google’s right to use what you send to improve its products. The paid tier starts the moment you link a billing account and prepay at least $5. It adds Gemini 4 Argon and Gemini 3.1 Pro Preview, higher limits, context caching and the Batch API, and your data stops being used for training.
Every price and limit here comes from Google’s AI Studio pricing, rate limits and billing pages and the Gemini API terms, checked on November 2, 2026. Several prices change on January 1, 2027, and two older models are rerouted on December 1, 2026, so we give both numbers. Current credit routes are on our Gemini API credits page.
What you get for free
The free tier means free input and output tokens, limited access to certain models, and Google AI Studio access. New accounts begin on it automatically; there’s nothing to activate beyond creating an API key.
The free models are Gemini 3.8 Flash, Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, Gemini 3.1 Flash-Lite, Gemini 3 Flash Preview, the deprecated Gemini 2.5 Flash, 2.5 Flash-Lite and 2.5 Pro, the Flash TTS models, Gemini Embedding and Embedding 2, the Robotics ER models and Gemma 4. Models with no free tier are Gemini 4 Argon, Gemini 3.1 Pro Preview, Gemini Omni Flash, every image generation model including the Nano Banana line, and Lyria music models. Context caching isn’t offered on the free tier for any model listed here.
Free tier rate limits
Rate limits are measured in requests per minute, input tokens per minute and requests per day. An RPM limit of 20 means the 21st request in a minute returns an error, and limits apply per project, not per API key, so extra keys buy you nothing. Daily quotas reset at midnight Pacific time.
Google no longer prints a per-model table of free tier RPM, TPM and RPD on the public page. Your active limits are shown in AI Studio and update automatically when your tier changes. Treat any figure that isn’t in your own AI Studio console as outdated.
Two limits are published. Grounding with Google Search on Gemini 3 and newer models defaults to 1,500 requests per day per project, and the Batch API has its own caps of 100 concurrent batch requests, a 2 GB input file and 20 GB of file storage, with batch enqueued token limits that start at Tier 1.
What the paid tier costs per million tokens
Prices are USD per one million tokens on the Standard tier. Output prices include thinking tokens.
| Model | Input | Output | Context caching | Free tier |
|---|---|---|---|---|
| Gemini 4 Argon | $2.00 ($4.00 from December 1, 2026) | $10.00 ($20.00 from December 1, 2026) | $0.10 ($0.20 from December 1, 2026), implicit only | No |
| Gemini 3.1 Pro Preview | $2.00 up to 200k tokens, $4.00 above | $12.00 up to 200k, $18.00 above | $0.20 or $0.40 plus $4.50 per 1M tokens per hour storage | No |
| Gemini 3.8 Flash | $0.75 ($1.50 from January 1, 2027) | $3.75 ($7.50 from January 1, 2027) | $0.075 ($0.15 from 2027) plus $0.50 per 1M tokens per hour | Yes |
| Gemini 3.6 Flash | Same as 3.8 Flash | Same as 3.8 Flash | Same as 3.8 Flash | Yes |
| Gemini 3.5 Flash-Lite | $0.30 | $2.50 | $0.03 plus $1.00 per 1M tokens per hour | Yes |
| Gemini 3.1 Flash-Lite | $0.25 text, image, video; $0.50 audio | $1.50 | $0.025 or $0.05 audio plus $1.00 per hour | Yes |
| Gemini 2.5 Flash (deprecated) | $0.30; $1.00 audio | $2.50 | $0.03; $0.10 audio | Yes |
| Gemini 2.5 Flash-Lite | $0.10; $0.30 audio | $0.40 | $0.01; $0.03 audio | Yes |
| Gemini 2.5 Pro (deprecated) | $1.25 up to 200k, $2.50 above | $10.00 up to 200k, $15.00 above | $0.125 or $0.25 plus $4.50 per hour | Yes |
| Gemini Embedding 001 | $0.15 | Not applicable | Not applicable | Yes |
The deprecations matter if you hard-code model names. From December 1, 2026, requests to gemini-2.5-flash route to Gemini 3.5 Flash-Lite and are billed at its rates, and requests to gemini-2.5-pro route to Gemini 3.8 Flash. Pin your model IDs and you won’t get a surprise.
As a working comparison, a million input tokens plus 200,000 output tokens costs $1.50 on Gemini 3.8 Flash today, $0.18 on 2.5 Flash-Lite and $4.00 on Gemini 4 Argon.
Context caching and the Batch API
Context caching charges a reduced rate for tokens you’ve already sent, plus an hourly storage fee. On Gemini 3.8 Flash cached tokens cost $0.075 per million against $0.75 for fresh input, a 90% reduction, with storage at $0.50 per million tokens per hour until 2027. On Gemini 3.1 Pro Preview the cached rate is $0.20 against $2.00, with storage at $4.50 per hour, so an idle cache can cost more than it saves. Gemini 4 Argon supports implicit caching only.
The Batch API halves every price in the table. Gemini 3.8 Flash is $0.375 in and $1.875 out through December 31, 2026, Gemini 3.5 Flash-Lite is $0.15 in and $1.25 out, and embeddings are $0.075. Batch is a paid feature with enqueued token limits per tier; for Gemini 3.8 Flash those are 3 million tokens at Tier 1, 400 million at Tier 2 and 1 billion at Tier 3. Anything that can wait overnight belongs here.
Grounding and tools
Grounding with Google Search on Gemini 3 and newer models includes 5,000 free requests per month shared across models, then $14 per 1,000, under the 1,500 per day default cap. On the 2.5 models it’s 1,500 requests per day free, then $35 per 1,000 grounded prompts. Only requests that return at least one web grounding URL are charged.
What Google does with your free tier data
Google’s price table has a column titled “Used to improve our products,” and for every free-tier model it reads “Yes” under the free tier and “No” under the paid tier. The Gemini API terms spell out what that means. For unpaid services, Google uses the content you submit and the responses generated to provide, improve and develop its products, and human reviewers may read, annotate and process your input and output. Google tells you not to submit sensitive, confidential or personal information to the unpaid services, and we’d take that literally.
Paid services carry the opposite commitment. Google doesn’t use your prompts or responses to improve its products and logs them only for a limited period. The API counts as paid only when accessed through a Cloud project with an active billing account, and users must be 18 or older. In the European Economic Area, Switzerland and the United Kingdom, the paid data terms apply even to AI Studio and unpaid quota, and only paid services may serve users there.
Usage tiers and how to turn on billing
The paid tier has three usage tiers, and your rate limits follow the tier.
- Free: any active project. Spend-based limit not applicable.
- Tier 1: link an active billing account. Monthly cap $250, with a $10 spend limit per rolling 10 minutes.
- Tier 2: $100 in cumulative payments plus 3 days since the first successful payment. Monthly cap $2,000, $200 per 10 minutes.
- Tier 3: $1,000 in cumulative payments plus 30 days. Monthly cap $20,000 to $100,000 or more, $200 per 10 minutes.
To enable billing, open the API keys or Projects page in AI Studio and click “Set up billing” on the project. New users choose a country, accept the terms and enter payment details, then pick Prepay or Postpay. Prepay has a $5 minimum and a $5,000 maximum, and API access stops when the balance hits $0. Postpay charges your card at month end or sooner if you reach your tier’s cap. Free to Tier 1 usually takes effect instantly, and later upgrades happen automatically within about 10 minutes.
The $300 Google Cloud trial credit can’t pay for Gemini API or AI Studio usage for accounts opened after March 2, 2026. If you want credits to cover Gemini, the route is the Google for Startups Cloud Program with Gemini through Vertex AI, covered in our startup credits guide, or the Google Cloud credits page for everything else.
Watching spend and going back to free
Usage appears in AI Studio under Dashboard, then Usage, and the same costs show in the Cloud Billing console next to your other Google Cloud charges. Monthly caps are enforced per billing account, and an experimental per-project cap lives on the AI Studio Spend page.
To return to the free tier, disable billing on each project you want to downgrade. Unused prepaid credit isn’t refunded when you do this; prepaid balances are refundable only when you switch to a Postpay account. Spend the balance down first.
How Gemini’s free tier compares with other API providers is on our free tiers page, and the Google Cloud free credits hub covers the cloud side.
The short version
- Flash, Flash-Lite, embeddings and TTS models are free with per-project rate limits; Argon, 3.1 Pro Preview and image models are paid only.
- Paid prices run from $0.10 per million input tokens on 2.5 Flash-Lite to $2.00 on Gemini 4 Argon, rising to $4.00 on December 1, 2026.
- Context caching cuts input cost by about 90% plus hourly storage, and Batch halves every price.
- Free tier prompts may be read by humans and used to improve Google’s products; paid tier prompts are not.
- Billing starts with a $5 prepay or a Postpay account in AI Studio, and the $300 cloud trial credit does not apply.
FAQ
Is Gemini API free to use
Yes, for the models marked free of charge, which include Gemini 3.8 Flash, 3.6 Flash, 3.5 Flash-Lite, 3.1 Flash-Lite and the embedding models. Rate limits apply per project, and your prompts can be used to improve Google’s products on the free tier.
How much does Gemini API cost per million tokens
On the Standard tier Gemini 3.8 Flash is $0.75 input and $3.75 output through December 31, 2026, Gemini 3.5 Flash-Lite is $0.30 and $2.50, Gemini 3.1 Pro Preview is $2.00 and $12.00 for prompts under 200k tokens, and Gemini 4 Argon is $2.00 and $10.00 until December 1, 2026, when it doubles.
What is the Gemini API free tier token limit
Google publishes the limit dimensions, requests per minute, input tokens per minute and requests per day, but the per-model figures are shown in your AI Studio rate limits view rather than on the public docs page. Limits apply per project and daily quotas reset at midnight Pacific time.
Does Gemini API free tier require a credit card
No. New accounts start on the free tier with an API key from AI Studio and no payment method. A billing account is required only for the paid tier, which has a $5 prepay minimum.
Does Google use my data on the Gemini API free tier
Yes. Free-tier usage is used to improve Google’s products, and human reviewers may read your input and output. Paid usage through a project with an active billing account is not used for product improvement.