iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Gemini API cost is not one fixed price. Your bill depends on the model, input and output tokens, cached tokens, tools, and service tier you select. Separately, every project is constrained by requests-per-minute (RPM), tokens-per-minute (TPM), requests-per-day (RPD), and rolling spend limits. The free tier has limited access and free token allowances; paid tiers require Cloud Billing but provide higher capacity and additional features. This guide explains how those systems interact, how to interpret a 429 RESOURCE_EXHAUSTED response, and what to check before increasing capacity or publishing a cost estimate.
What determines the Gemini API price?
Gemini pricing is model-specific rather than universal. Google publishes separate rows for input tokens, output tokens, cached tokens and storage, tool or grounding usage, and service tiers such as Standard, Batch, Flex, and Priority. You must read the row for the exact model and tier you plan to call; multiplying a headline token price across every modality can materially misstate your bill.
Input and output tokens are billed separately
Input tokens are the text, images, audio, video, documents, and other content sent to the model, according to the modalities supported by that model. Output tokens are the generated response. Some model rows explicitly include thinking tokens in the output price, so check that wording before estimating an agent or reasoning-heavy workflow.
Caching and tools add their own line items
Context caching can charge for cached tokens and for storage. Grounding and other tools have separate pricing rows. A request that looks inexpensive from its input and output token counts can therefore cost more when it uses a tool or keeps a large cache active.
Service tier changes the economics
Standard, Batch, Flex, and Priority are separate service choices in the pricing table. Batch API is advertised as a 50% cost reduction on paid tiers, which can suit asynchronous jobs that do not need an immediate response. Confirm that the model supports the chosen tier and that the reduction applies to the workload you are submitting.
How much does Gemini 4 Argon Standard cost?
The following is a dated example from Google’s pricing table. Gemini 4 Argon Standard input is listed at $2.00 per 1 million tokens through November 30, 2026, then $4.00 from December 1, 2026. Output is listed at $10.00 per 1 million tokens through November 30, then $20.00 from December 1. These are effective-date prices, not a permanent Gemini-wide rate.
| Model and tier | Usage | Price through Nov. 30, 2026 | Price from Dec. 1, 2026 |
|---|---|---|---|
| Gemini 4 Argon Standard | Input tokens | $2.00 per 1 million tokens | $4.00 per 1 million tokens |
| Gemini 4 Argon Standard | Output tokens, including thinking tokens where the row specifies them | $10.00 per 1 million tokens | $20.00 per 1 million tokens |
For an illustration using the first price period, 100,000 input tokens would be $0.20 and 50,000 output tokens would be $0.50 before any cache, tool, or other charges. The arithmetic changes when the effective date, model, modality, or service tier changes. Recheck the live model table immediately before committing a budget because Google lists future effective dates and can revise model availability or prices.
What is included in the free, paid, and enterprise tiers?
Tier selection affects both capacity and data handling, not just the invoice. Free access offers limited model availability plus free input and output tokens. Paid access raises rate limits, adds context caching and Batch API access, and unlocks advanced models. Google states that prompts and responses in the paid tier are not used to improve its products. Enterprise adds support, security and compliance features, provisioned throughput, and volume discounts.
| Tier | Capacity and model access | Features and data treatment |
|---|---|---|
| Free | Limited model access and limited quotas | Free input and output token allowance |
| Paid | Higher rate limits and advanced model access | Context caching and Batch API access; paid prompts and responses are not used to improve Google products |
| Enterprise | Provisioned throughput and capacity for organizational workloads | Support, security and compliance features, and volume discounts |
“Paid” does not mean unlimited. A paid project still has RPM, TPM, RPD, and spend-based ceilings. Enterprise capacity and discounts are also subject to the agreement and model availability you receive.
What billing qualifications apply to each paid tier?
Billing qualification is separate from per-request quota. The thresholds below describe Google’s published progression requirements and listed billing-account caps; they are not a promise that every model will expose the same RPM, TPM, or RPD values.
| Tier | Qualification | Listed billing-account cap |
|---|---|---|
| Tier 1 | Active billing account | $250 |
| Tier 2 | $100 paid and three days elapsed from the first successful payment | $2,000 |
| Tier 3 | $1,000 paid and 30 days elapsed from the first successful payment | $20,000–$100,000+ |
To upgrade to the paid tier, Google’s getting-started documentation specifies a minimum $5 prepaid credit and requires Cloud Billing. Payment history and elapsed time matter for the higher tiers, so adding a billing account alone does not immediately grant Tier 2 or Tier 3 qualification.
What do RPM, TPM, and RPD mean?
Gemini quotas measure different kinds of pressure on a project. RPM limits request volume, TPM limits token throughput, and RPD limits the number of requests over a day. Google applies these limits per project, not per API key, and RPD resets at midnight Pacific time.
RPM: requests per minute
RPM counts calls during a rolling or defined one-minute window. A small prompt can still fail if many clients send requests at once. Spreading calls over time, queueing bursts, and avoiding duplicate retries reduce RPM pressure without changing your token bill.
TPM: tokens per minute
TPM measures token throughput. A handful of very large prompts can exhaust TPM even when RPM is low. Trim repeated instructions, cap unnecessary history, use caching where it is economical, and schedule large jobs so their token load is distributed across minutes.
RPD: requests per day
RPD limits daily request count. It is independent of token size, so many short calls can reach RPD while plenty of TPM remains. The quota resets at midnight Pacific time; do not assume that reset occurs at midnight in your local time zone.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Clear out junk files and repair common Windows errors3Fix the driver behind crashes, sound loss and screen glitchesWhy model choice changes the limits
Limits vary by model and tier. Experimental and preview models generally have tighter limits than established production models. Switching models can therefore change your RPM, TPM, RPD, context window, price, and eligibility at the same time.
Why am I getting a 429 RESOURCE_EXHAUSTED error?
A 429 RESOURCE_EXHAUSTED response means the request hit a limit. The cause may be RPM, TPM, RPD, or a rolling spend ceiling; the correct fix depends on which dimension is exhausted.
Separate transient rate pressure from a daily limit
If a short burst causes failures, reduce concurrency and retry with exponential backoff and jitter. A backoff policy should stop retrying after a bounded number of attempts and preserve the original request so that retries do not create a second burst. If the project has exhausted RPD, waiting for the midnight-Pacific reset is required; backoff alone cannot create new daily quota.
Check rolling spend limits
Google documents rolling 10-minute spend-rate limits of $10 for Tier 1, $50 for Tier 2, and $200 for Tier 3. The free tier has no listed spend-based value. If a spend-based limit is reached, the API returns 429 RESOURCE_EXHAUSTED even when RPM or TPM appears available. Reduce the amount of paid usage in that ten-minute window, queue work, or move to a tier with a higher published ceiling if your project qualifies.
The Tool Desk
Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Inspect the project, not just the API key
Because limits are project-level, creating additional API keys in the same project does not multiply the quota. Confirm which project each key targets, then inspect that project’s active limits in AI Studio. Preview and experimental model rows can have stricter values than the model you used for an earlier test.
How can I increase Gemini API quota?
Quota increases come from selecting an appropriate model and tier, attaching billing, and meeting the published payment qualifications; they are not obtained by rotating keys. Use this sequence before changing application code:
- Identify the exact model and service tier. Record whether the call uses Standard, Batch, Flex, or Priority and whether the model is preview or experimental.
- Open the project’s limits in AI Studio. Check the current RPM, TPM, and RPD values for that model rather than relying on a limit shown for another model.
- Attach Cloud Billing when moving off free access. Paid access requires billing and a minimum $5 prepaid credit according to Google’s getting-started guidance.
- Meet the higher-tier qualification. Tier 2 requires $100 paid plus three days from the first successful payment; Tier 3 requires $1,000 paid plus 30 days.
- Recheck spend ceilings and billing caps. Higher capacity does not remove the rolling 10-minute spend guard or the billing-account cap.
- Change the workload shape. Queue bursts, lower concurrency, reduce repeated context, and use Batch for eligible asynchronous work before requesting more capacity.
There is no single quota number to request for “Gemini.” Limits are model-, project-, and tier-dependent, so document the exact combination when opening a capacity request or troubleshooting a production deployment.
How should I estimate a Gemini API budget?
Estimate each charge separately and then compare the result with the project’s spend safeguards. A practical worksheet has these lines:
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallCrashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minute- Input tokens × the model’s input price per million tokens.
- Output tokens, including thinking tokens when the model row says they are billable, × the output price per million tokens.
- Cached-token usage and cache storage, using the model’s caching rows.
- Grounding and other tool calls, using their separate pricing rows.
- Service-tier adjustments, including the advertised 50% Batch reduction where it applies.
Keep traffic volume and spending in different calculations. RPM, TPM, and RPD tell you whether requests can be admitted; token prices tell you what admitted usage costs. A workload can fit under RPM but exceed TPM, or remain under both while crossing a rolling spend limit.
Model a range, not a single point estimate
Use low, expected, and high token counts for prompts and responses. Include retries, long-context outliers, tool calls, and cache storage. For a model with a scheduled price change, calculate both the current and future effective-date totals so a deployment that crosses the date is not underfunded.
Watch usage continuously
Monitor prepaid balance, project and billing-account spend caps, and the Dashboard > Usage view. Alert on rising TPM and spend before the API starts returning 429 responses. Keep a record of model name, tier, effective date, and token categories with each estimate.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What should I verify before publishing a price or quota figure?
- Confirm the model name and whether it is stable, preview, or experimental.
- Read the exact input, output, caching, storage, and tool rows for the selected service tier.
- Check whether thinking tokens are included in the output definition.
- Record the effective date, region or account qualification if the table specifies one, and whether a price is promotional or scheduled to change.
- Check the project’s live RPM, TPM, and RPD values in AI Studio.
- State that RPD resets at midnight Pacific time and that limits apply per project, not per API key.
- Include the applicable rolling 10-minute spend limit and billing-account cap.
Pricing pages and model catalogs are date-sensitive. Recheck them immediately before publication or deployment rather than carrying forward a number from an earlier estimate.
Free tools Windows power users keep installed
One-click scans. No signup required.
Gemini API pricing and quota FAQ
Is the $5 prepaid credit a recurring monthly fee?
No. The documented $5 figure is the minimum prepaid credit used to upgrade to paid access. It is not described as a monthly subscription charge; ongoing cost depends on actual model and feature usage.
Can I use Batch API on the free tier?
Batch API access is listed as a paid-tier feature. If asynchronous processing is important, budget for paid access and verify that the selected model supports Batch before designing the workflow around its advertised reduction.
Does a higher billing cap automatically raise every quota?
No. Billing-account caps and request/token quotas are separate controls. A project can qualify for a higher billing cap while a particular preview model still has tighter RPM, TPM, or RPD limits.
Are tool calls included in the model’s token price?
Not necessarily. Grounding and other tools have their own pricing rows, so include those charges as separate estimate lines instead of assuming they are covered by input or output tokens.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Frequently Asked Questions
Is the $5 prepaid credit a recurring monthly fee?
No. The documented $5 figure is the minimum prepaid credit used to upgrade to paid access. It is not described as a monthly subscription charge; ongoing cost depends on actual model and feature usage.
Can I use Batch API on the free tier?
Batch API access is listed as a paid-tier feature. If asynchronous processing is important, budget for paid access and verify that the selected model supports Batch before designing the workflow around its advertised reduction.
Does a higher billing cap automatically raise every quota?
No. Billing-account caps and request/token quotas are separate controls. A project can qualify for a higher billing cap while a particular preview model still has tighter RPM, TPM, or RPD limits.
Are tool calls included in the model’s token price?
Not necessarily. Grounding and other tools have their own pricing rows, so include those charges as separate estimate lines instead of assuming they are covered by input or output tokens.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
