Short answer: Google offers a free tier for the Gemini Developer API, but access is limited to certain models and each project’s limits depend on the model and usage tier. Google’s general pricing description does not name the eligible models. Its current model guidance recommends gemini-3.5-flash-lite and gemini-3.8-flash for new projects, but those recommendations alone do not establish that either model is free. Check the model’s free-tier status on Google’s pricing page and its active limits in AI Studio before planning around free usage. (aistudio.google.com)
This guide covers the Gemini Developer API and Google AI Studio. Do not assume its limits apply to Vertex AI.
Which Gemini API models have free access?
Google describes free access as limited to certain models, with free input and output tokens up to the applicable limits. Eligibility is model-specific. The official sources cited here establish the following, but do not establish a positive free-tier choice for every model listed. (aistudio.google.com)
| Model | What the official documentation establishes about free access |
|---|---|
Gemini 3.5 Flash-Lite (gemini-3.5-flash-lite) |
Google recommends it for new projects and publishes this exact model ID. The model page alone does not confirm free-tier eligibility. Check its current pricing row and your project’s active limits. (ai.google.dev) |
Gemini 3.8 Flash (gemini-3.8-flash) |
Google recommends it for new projects and publishes this exact model ID. The model page alone does not confirm free-tier eligibility. Check its current pricing row and your project’s active limits. (ai.google.dev) |
Gemini 4 Argon (gemini-4-argon) |
Google’s pricing page explicitly says it is not available on the Free tier. (aistudio.google.com) |
| Gemini 2.5 models | Google says access is limited to users who have actively used these models in the past. That restriction does not by itself establish whether a particular project has free-tier access. Check the model’s pricing status and your project. (ai.google.dev) |
Important: A model being recommended for new projects or available in AI Studio does not prove it is included in the free API tier. Confirm free eligibility on the pricing page before using it as a budget assumption. The official rate-limit guidance directs users to check active limits in AI Studio rather than giving one universal quota for every model. (aistudio.google.com)
What do the quota numbers mean?
Google says Gemini API limits are usually measured in three dimensions:
- RPM: requests per minute.
- TPM: input tokens per minute.
- RPD: requests per day.
Each applicable limit is enforced independently: exceeding one can trigger a rate-limit error even if the others have not been reached. Limits vary by model; additional dimensions may apply to certain models or tools. Daily request limits reset at midnight Pacific Time. (ai.google.dev)
Where to check the limits for your project
- In Google AI Studio, select the project associated with the API key your application uses.
- Open View your active rate limits.
- Check the exact model identifier and the limits displayed for that project.
- Compare each limit with expected peak-minute requests, input tokens per minute, and daily requests. Check for any additional model- or tool-specific limits.
Google applies rate limits per project, not per API key. Creating another key for the same project does not create a separate quota. (ai.google.dev)
Can the free tier support production?
It depends on your workload, project limits, and model eligibility. Google positions the free tier for developers and small projects, and the paid tier for applications needing higher volumes, higher rate limits, or access to more models. That is guidance, not a rule that every production application must pay. Compare measured peak and daily use with your project’s active limits, allowing headroom for bursts and retries. (aistudio.google.com)
Google says exceeding an applicable rate limit triggers an error; do not assume an over-limit free-tier request automatically switches to paid billing. Google describes the paid tier as providing higher rate limits and access to its most advanced models. (ai.google.dev)
One production consideration is data use: Google’s pricing page says content submitted on the free tier is used to improve its products, while paid-tier content is not used for that purpose. Review the applicable terms before sending sensitive data. (aistudio.google.com)
An AI Studio subscription is not the same as API billing. Google says Google AI plan benefits for developer use apply only in the AI Studio web interface; direct Gemini API use through an API key or external application is managed and billed separately.
Budgeting if free limits are not enough
Estimate token costs by calculating input and output separately:
Estimated token cost = (input tokens ÷ 1,000,000 × input price per million) + (output tokens ÷ 1,000,000 × output price per million)
Google lists introductory Gemini 3.8 Flash rates through December 31, 2026 of $0.75 per million input tokens and $3.75 per million output tokens. Standard rates are scheduled to take effect January 1, 2027: $1.50 input and $7.50 output per million tokens. (aistudio.google.com)
Illustrative monthly estimate: Suppose an application makes 100,000 requests per month, averaging 2,000 input tokens and 500 output tokens per request. That is 200 million input tokens and 50 million output tokens. At the introductory rates:
- Input: 200 × $0.75 = $150
- Output: 50 × $3.75 = $187.50
- Estimated total: $337.50 per month
At the scheduled January 1, 2027 standard rates, the same token volumes would cost $675. This is an estimate, not a quote: it assumes every request uses this model and these average token counts, and excludes separately priced features or services. Recalculate using the model, measured usage, and rates that apply to your project.
Bottom line
Gemini API free access is model-specific; there is no single quota that applies to every model or project. Confirm a model’s free-tier status on Google’s pricing page, then use View your active rate limits in AI Studio to check your project’s quotas. Limits are project-scoped, daily request limits reset at midnight Pacific Time, and any one applicable limit can constrain usage. If your workload needs more capacity or a paid-only model, estimate costs using current input and output rates and confirm billing is configured for the project.
Sources
Gemini Developer API pricing | Google AI Studio
Rate limits | Gemini API | Google AI for Developers
Gemini 3.5 Flash-Lite | Gemini API | Google AI for Developers
Gemini 3.8 Flash | Gemini API | Google AI for Developers
Gemini deprecations | Gemini API | Google AI for Developers
What's new in Gemini 3.8 Flash | Gemini API | Google AI Studio