Short answer
For a personal Gemini account without an AI plan, Google describes access as subject to “standard limits,” not a fixed number of free prompts per day. App usage is compute-based: prompt complexity, model and feature choice, and chat length affect how much of your limit you use. Limits refresh every five hours until you reach the weekly limit, but the cited guidance does not state a fixed free-account allowance or a universal weekly reset time. (support.google.com)
The Gemini API has separate free access and limits. Free access applies only to certain models; rate limits depend on the model and project. App prompts and API quotas are not interchangeable. (aistudio.google.com)
Gemini app: what the free limit means
Google lists “standard limits” for personal accounts without an AI plan and warns that app limits and feature availability may change. With automatic model selection enabled, Gemini chooses a model for each prompt; the available models and access depend on Google’s current app guidance. (support.google.com)
Google says the app notifies you when you are close to a limit and, when you reach it, tells you when the limit will refresh. The general limit refreshes every five hours, subject to the weekly limit. The cited guidance does not establish that a smaller model or other features will remain available after either limit. Follow the options and refresh time shown in your app rather than assuming a particular fallback or weekly reset. (support.google.com)
Some features have separate constraints. Deep Research, for example, has daily request and concurrent-request limits. The cited guidance does not establish a current numerical free allowance for Deep Research or a universal free image-generation quota. (support.google.com)
Gemini API: free access is not an unlimited production quota
Google’s API Free Tier includes free input and output tokens for certain models, but free eligibility is model-specific. Gemini 3.5 Flash-Lite (gemini-3.5-flash-lite) is a candidate to assess for text extraction: Google describes it as designed for high-throughput tasks, document parsing, and simple data extraction. However, the retrieved pricing-page information does not confirm that this model is Free Tier-eligible. Check the model’s current pricing row before treating it as free. (aistudio.google.com)
Google says API rate limits are measured using dimensions such as requests per minute (RPM), input tokens per minute (TPM), and requests per day (RPD). They vary by model and apply per project, not per API key. Google directs developers to View your active rate limits in AI Studio; daily request quotas reset at midnight Pacific time. The guidance retrieved here does not provide numerical free-tier RPM, TPM, or RPD values for the model and project in the example below. (aistudio.google.com)
Illustrative capacity check—not a Google-published quota: Suppose a text-extraction service expects 10,000 requests per day, peaks at 20 requests per minute, and averages 1,500 input tokens and 250 output tokens per request. That is about 15 million input tokens and 2.5 million output tokens per day. At the assumed average input size, the peak also requires 20 RPM and 30,000 input TPM. Compare those requirements with the active RPM, input TPM, and RPD values shown in AI Studio for the chosen model and project. Because the retrieved documentation does not establish this model’s Free Tier eligibility or this project’s numerical limits, this example cannot determine whether the workload fits within free access. If the active limits are too low, reduce or queue traffic, evaluate another model whose Free Tier eligibility you confirm, or plan for paid use. (aistudio.google.com)
A rate-limit error does not by itself mean the project has switched to paid billing. Google says moving to the Paid Tier requires linking a billing account and prepaying at least $5 (or the equivalent in another currency). Check the project’s billing status before sending production traffic. (ai.google.dev)
Google’s API pricing page distinguishes Free Tier content as used to improve Google products from Paid Tier content as not used for that purpose. These are the page’s tier-level statements; confirm the terms applicable to your account and project before using them as a deployment criterion. (aistudio.google.com)
Example: estimate a paid API token bill
API cost depends on the API model and token volume—not on your free-app prompt allowance. Suppose a workload processes 10 million input tokens and 5 million output tokens using Gemini 4 Argon through the API. Google lists an introductory rate of $2 per 1 million input tokens and $10 per 1 million output tokens through November 30, 2026:
- Input: 10 × $2 = $20
- Output: 5 × $10 = $50
- Estimated token charge: $70 USD
Google lists rates of $4 per 1 million input tokens and $20 per 1 million output tokens starting December 1, 2026. The same hypothetical volume would then cost $140 USD. Argon is not available on the API Free Tier, so this is a paid-rate illustration, not a free-tier estimate. It excludes separately priced tools or other charges; actual costs depend on measured usage and applicable model rates. (aistudio.google.com)
Which option should you choose?
- Use the free Gemini app for personal, interactive tasks if its changing standard limits and model access fit your needs. It is not a fixed daily message allowance or an API budget.
- Evaluate the API Free Tier only after confirming a specific model’s eligibility and checking the project’s active RPM, input TPM, and RPD limits in AI Studio. Do not assume the extraction example—or Gemini 3.5 Flash-Lite—qualifies for free use without checking.
- Plan paid API billing if the workload needs higher limits, paid-only models or features, or the Paid Tier’s data-use terms. Linking billing is separate; a consumer Google AI subscription does not pay for direct API calls. (
ai.google.dev)
Bottom line: The cited guidance gives no fixed daily prompt count for the personal Gemini app. API free access is separate, and its model eligibility and project-specific limits must be checked before using it for production planning.
Sources
Gemini Apps limits & feature availability — Google Help
Understanding Gemini Apps Usage Limits — Google Help Community
Gemini Developer API pricing — Google AI Studio
Gemini 3.5 Flash-Lite — Google AI for Developers