Google Maps Geocoding API Pricing 2026: Real Monthly Costs

How Google Maps Geocoding API pricing works in 2026

Geocoding converts an address into coordinates (forward geocoding) or coordinates into an address (reverse geocoding). Google bills it as a pay-as-you-go SKU in its Essentials category, priced per 1,000 requests.

Two changes since March 2025 shape every calculation below:

• No pooled credit. Google replaced the single $200 monthly credit with a free usage cap per SKU. Essentials SKUs such as Geocoding get 10,000 free events a month. Free usage on one API no longer absorbs paid usage on another.
• Volume tiers per SKU. The price per 1,000 drops as monthly volume crosses each tier boundary, and each tier's rate applies to the requests inside that tier.

Source: Google Maps Platform core services pricing. Rates in USD, excluding tax and any negotiated discount.

Google also sells monthly subscription plans that bundle Essentials and Pro calls. They change the effective rate if your usage matches the plan, and they do not cover Enterprise SKUs. Our full Google Maps API pricing breakdown models those plans for a complete store locator, so this article stays on the geocoding line item.

What geocoding costs at 10K, 100K and 1M requests a month

Here is the monthly bill at five volumes, applying each tier's rate only to the requests inside that tier. The Woosmap column uses the Localities Geocode rates from the Woosmap pricing page, checked on 2026-10-08: 10,000 free requests a month, then $2.04 per 1,000 up to 100,000, $1.64 up to 500,000 and $1.53 beyond. The Woosmap amounts are reproducible estimates from these published rates, not quotes, invoices or guaranteed savings. They assume the first 10,000 requests each month are free and that each rate applies only to the requests inside its tier (graduated tiers, as on Google), which is the method the public cost calculator on that page uses. We keep cents in the arithmetic; the calculator rounds the running total down after each tier, so it shows slightly lower figures (for example $81 rather than $81.60 at 50,000 requests). How your contract applies the tiers can change these totals, so treat them as a forecast, not a quote.

Worked example at 500,000 requests on Google: the first 10,000 are free, the next 90,000 cost $450 at $5.00, and the next 400,000 cost $1,600 at $4.00, for $2,050 a month or $24,600 a year.

Woosmap assumptions (rates checked on 2026-10-08): 10,000 free requests a month, graduated tiers. This article does not verify how a Woosmap contract applies the tiers. If your contract bills the whole monthly volume at the rate of the tier it reaches instead, the 500,000 and 1,000,000 rows change. Confirm the billing calculation with Woosmap before you put these figures in a budget.

Three things this table tells a budget owner:

1. Below 10,000 requests a month, geocoding is free on both. If that is your steady state, the decision is about terms and data handling, not price.
2. Between 50,000 and 500,000 requests, the gap is a stable ratio. On the graduated-tier estimate above, Woosmap lands at about 41% of the Google line item, so estimated annual savings scale linearly: about $3,200 a year at 100,000 requests and about $14,500 at 500,000. These are estimates from list rates, not guaranteed savings.
3. Above 1 million requests, the gap closes. Google's list rate drops to $1.50 per 1,000 between 1 million and 5 million requests, close to Woosmap's $1.53, and to $0.38 above 5 million. At that scale list prices stop being the comparison: model your own tier mix and ask both vendors for a volume quote.

If these numbers already make the case, put your current invoice next to the table and check which tier your monthly volume actually lands in.

The billing trap: geocoding after autocomplete

Most address flows do not start with a typed full address. They start with an autocomplete field, and the user picks a suggestion. How you turn that pick into coordinates decides what the whole session costs, and the geocoding line item on your invoice hides most of it.

Google bills autocomplete in sessions. A session starts with Autocomplete (New) and a session token, and it ends with a Place Details (New) call or an Address Validation call. What happens next depends on how it ends:

• Ended with Place Details Pro or Enterprise: every autocomplete request in the session is free. You pay for the Details call: $17 per 1,000 at Pro, $20 at Enterprise.
• Ended with Place Details Essentials: the Details call costs $5 per 1,000, but the first 12 autocomplete requests of the session are billed ($2.83 per 1,000 in the 10K - 100K tier). Requests 13 and above are free.
• Not ended at all: if the session never reaches a Details or Address Validation call, autocomplete falls back to per-request billing.

A Geocoding API call does not close a session. So a common "cheap" pattern - autocomplete for suggestions, then the Geocoding API at $5 per 1,000 to get coordinates for the chosen text - leaves every session abandoned, and each autocomplete request is billed on its own.

The less obvious result is the Essentials termination. It looks like the budget option, $5 instead of $17. But each billed autocomplete request adds $2.83 per 1,000 sessions. The break-even sits at about four requests:

Rates from the 10K - 100K tier; free usage caps excluded.

If your users usually type five characters or more before picking a suggestion, the "cheaper" Details tier costs more per session than the Pro tier. Before you forecast geocoding spend, count autocomplete requests per session in your own logs. That one number decides which Google configuration is cheapest, and it is the number most pricing comparisons skip.

Woosmap Localities works differently: autocomplete does not require a session token, and no terminating call is needed to keep suggestion requests from being billed per session. Teams that resolve the picked suggestion with a Geocode call pay the Geocode rate in the table above, with no session state to manage.

Google Geocoding API vs the alternatives

Google is not the only option, and the cheapest option is not always the right one. This table compares what matters for a geocoding workload.

Sources: Google and Woosmap pricing pages as linked above; Nominatim usage policy; Mapbox pricing page and Geocoding API documentation (temporary and permanent geocoding rates, batch limit), read on 2026-10-07; HERE pricing page, Geocoding & Search Base plan, read on 2026-10-07. HERE shows its rates in euros and labels them indicative estimates, with the final price set by contract; we have not converted them to dollars. Google and Mapbox rates are in USD.

Three notes. Mapbox sells two geocoding modes: temporary geocoding is the cheaper one ($0.75 per 1,000 after 100,000 free requests), but its results may not be stored, while permanent geocoding ($5.00 per 1,000, no free tier) costs as much as Google's first paid tier and lets you keep the coordinates. HERE bills geocoding per 1,000 transactions after 30,000 free a month (EUR 0.70 per 1,000 up to 5 million on the Base plan, an indicative euro figure), and its Base plan does not include permanent geocoding. With both, check whether the rate you compare against Google allows the storage and display your architecture needs - a temporary rate you cannot cache from is not a like-for-like price. Mapbox's batch endpoint (up to 1,000 queries per call) matters if you geocode stored datasets rather than live input. And the public Nominatim service is a good fit for prototypes, but its usage policy caps you at one request per second and forbids heavy use, so production traffic means self-hosting it - which moves the cost from invoices to infrastructure and engineering time. For a provider-by-provider view, see our guides to Mapbox alternatives and Azure Maps alternatives.

How to choose: a decision table

Find your situation in one row.

If you sit in the third or fourth row, rebuild the cost table above with your own volumes before you scope a migration: the per-1,000 rates and free caps are all public.