How Much Does a Google Maps API Cost in 2026? Real Prices per 1,000 Places

Every "cheapest Google Maps API" article I have read compares a monthly subscription price against a per-call price and calls one of them cheaper. That is not pricing. It is marketing.

So I priced the same workload — one thousand delivered place records — on eight providers: the official Google Places API, DataForSEO, ScrapingDog, SerpApi, Apify, Bright Data, Outscraper, and my own Serpent Maps. The spread is wild: $0.006 to $18.60 per 1,000 places, a factor of three thousand. But the spread is not the story. The cheap end is selling a different object, and once you line the field lists up next to each other the comparison stops being about price at all. This is the same exercise I ran for general SERP API pricing, applied to local business data.

How I Priced the Same Workload

One workload, eight providers, one unit. The workload: deliver 1,000 place records for a local business search. I logged the billing unit (per call, per record, per task, or subscription search), the vendor's lowest published rate rather than their entry rate, and what one thousand places actually cost once the unit is normalised. Where a vendor sells places in pages, I divided by the page size their own documentation states. Where a provider returns a mix of ranked records and fully-enriched ones, I report both units rather than quietly picking the flattering one.

Two normalisations do most of the work in the table below. The official Places API caps maxResultCount at 20 places per search call, so its $32.00 per 1,000 calls is $1.60 per 1,000 places, not $32. And a DataForSEO task is one SERP of up to 100 results at the default depth, so its $0.60 per 1,000 tasks is $0.006 per 1,000 places. Both of those corrections cut against the version of this article I published in August, which quoted $32–$34 for Google and ~$5.00 for DataForSEO.

Sorted by the bolded per-1,000-places figure, which is each vendor's lowest published rate for the record described in the last column. Read on each vendor's own page on 8 September 2026; sources at the foot of the article. "Not field-compared" means I priced the row but did not audit its field list against ours, so treat those two as price-only entries.

Why Per-Unit Pricing Is Confusing

Every Google Maps API prices differently, and the units are not interchangeable:

• Per call or per request. You pay once and receive a page of places. Serpent Maps, ScrapingDog and the official Places API price this way. The cost per place depends entirely on how many places one call returns — 20 for Google and ScrapingDog, up to 100 for a Serpent Deep call.
• Per task. DataForSEO bills one task per SERP of up to 100 results, so a task is a page-priced unit with a much larger page.
• Per 1,000 records. You pay per place delivered. Apify actors, Outscraper and Bright Data price this way. Predictable, but enrichment is usually a separate line item.
• Subscription searches. You buy a monthly search allowance. SerpApi prices this way; a Google Maps search there returns up to 20 local results.

Convert everything to "cost per 1,000 places" and the ranking rearranges completely. A $32.00 per-1,000-calls rate that returns 20 places a call is $1.60 per 1,000 places. A $0.60 per-1,000-tasks rate that returns 100 results a task is $0.006. Those two headline numbers differ by 53×; the real per-place gap is 267×. The unit has to be stated as carefully as the rate — and then, separately, so does the record.

1. Official Google Places API — ~$6.60 to ~$18.60 per 1,000 Businesses

This is the figure I got most wrong in the August version, and I got it wrong in the direction that flattered me. Google's current pricing page splits the Places SKUs across tiers. Nearby Search and Text Search bill $32.00 per 1,000 calls on Pro ($35.00 Enterprise, $40.00 Enterprise + Atmosphere; neither SKU is offered on Essentials). Place Details bills $5.00 per 1,000 on Essentials, $17.00 on Pro, $20.00 on Enterprise and $25.00 on Enterprise + Atmosphere.

The old article read $32 per 1,000 calls as $32 per 1,000 businesses. It is not, because maxResultCount on Nearby Search accepts 1 to 20, default 20 — one search call returns up to twenty places. So the search step is $32.00 ÷ 20 = $1.60 per 1,000 places, and Place Details is charged once per place on top. A search-plus-details pipeline is therefore about $6.60 per 1,000 businesses with Essentials details, or about $18.60 with Pro details. Roughly half to a fifth of what I previously published.

Google also publishes real volume steps, and they are steep. Above 5,000,000 calls a month the marginal rates drop to $2.40 per 1,000 searches and $0.38 per 1,000 Essentials details — about $0.50 per 1,000 businesses, which undercuts every fully-enriched price in this article including ours, though it is still well above the SERP-row rates and above Serpent's ranked-list rate. That is a marginal rate at a volume almost nobody reaches, and it still requires a Google Cloud project, keys on every request, and field masks that bill by field group. But it is published, so it belongs in the table.

2. DataForSEO Google Maps SERP API — ~$0.006 per 1,000 Places

The second correction, and the largest. DataForSEO's Google Maps SERP API costs $0.60 per 1,000 tasks on the standard queue — $1.20 priority, $2.00 live. My August figure of "~$5.00 per 1,000 SERP pages" was wrong by more than 8× on their own unit.

It gets cheaper still once the unit is normalised. Their documentation sets depth to a default of 100 (maximum 700) and bills "per each SERP containing up to 100 results". One task is therefore up to a hundred places, which puts DataForSEO at roughly $0.006 per 1,000 places. Only ScrapingDog is anywhere near it; everything else in this comparison, us included, is at least sixteen times more per place. What you get for it is discussed below, and it is the whole reason this article no longer ends with a "we are cheapest" verdict.

3. ScrapingDog — ~$0.012 to ~$0.05 per 1,000 Places

Absent from the August comparison entirely. ScrapingDog's Google Maps API page publishes per-request rates directly: $1.00 per 1,000 requests on Lite, $0.45 on Standard, $0.33 on Pro and $0.24 on Premium with annual billing. The Lite plan is $40 a month for 200,000 credits and a Maps request costs 5 credits, which is the same $1.00 per 1,000 requests by a different route. Their pagination increments by 20, so a request is a page of twenty places: about $0.05 per 1,000 places at the entry plan and $0.012 at their lowest published rate.

4. SerpApi — $1.25 down to ~$0.098 per 1,000 Places

SerpApi sells a monthly search subscription, not pay-as-you-go, and Maps is a search type on the same meter rather than a separate product. Starter is $25 a month for 1,000 searches, and their Maps documentation paginates in steps of 20, so a search is twenty places: $1.25 per 1,000 places at the entry plan. Their largest published Cloud plan works out near $1.96 per 1,000 searches, or roughly $0.098 per 1,000 places, at $106,050 a month.

The August version of this article said unused SerpApi credits expire at month-end, in the body, in the decision list, in the FAQ and in this page's structured data. I could not source that claim, so I have removed it from all four places. Their pricing page as of 8 September 2026 makes no such statement. What it does say is that on a downgrade "any remaining searches from your previous plan will be added to your 'Extra Credits' balance", and that if you cancel before your plan expires you can keep using remaining searches until the renewal date. A negative claim about a competitor that I cannot point at on their own page has no business being on this site, and it had been sitting in schema markup where Google could surface it as a rich result.

5. Apify — $0.65 to $0.75 per 1,000 Fully-Enriched Records

The August version said "Apify Maps actors $1.00 to $5.00 per 1,000 records", which is a range across an actor marketplace rather than a price. Apify is a marketplace, so the honest thing is to name the actor. The one that matters for this comparison is scraperlink/google-maps-scraper, because it is the only cheap option I found that sells a record comparable to ours.

Its published rates are $0.50 per 1,000 results on the free and Starter plans, $0.45 on Scale ($199/mo) and $0.40 on Business ($999/mo). Popular times is an add-on at +$0.25 per 1,000 places, flat across all plans, and full reviews are another +$0.25 per 1,000 review records actually returned. So a fully-enriched place — base plus popular times — is $0.75 per 1,000 on the free and Starter plans, $0.70 on Scale, and $0.65 on Business. That is the like-for-like number, and it is the one I compare us against below.

6. Outscraper — $1.00 to $3.00 per 1,000 Records

Also missing from the August comparison. Outscraper's Google Maps pricing is volume-tiered per record: the first 500 places are free, then $3.00 per 1,000 from 501 to 100,000, then $1.00 per 1,000 above 100,000. I priced the row but did not audit its field list against ours, so treat it as a price-only entry.

7. Bright Data — $1.30 to $1.50 per 1,000 Records

The third addition. Bright Data's Google Maps scraper is $1.50 per 1,000 records pay-as-you-go, dropping to $1.30 per 1,000 on the $499-a-month Scale plan, which includes 384,000 records. The free tier covers 5,000 records a month. Same caveat as Outscraper: price verified, field list not compared.

8. Serpent Maps — $0.105 to $0.75 per 1,000 Places

Disclosure: I run Serpent API, so I will state this plainly and let the numbers stand next to everyone else's. Serpent's Google Maps API is per call, with four endpoints:

Prices are per 1,000 calls. Growth and Scale tiers discount against Default.

Because the rate is per call rather than per record, the per-place cost depends on how full the response is. On Deep there is one more thing to know, and it is the single most important line in this section: Deep has two flat price bands and a cliff between them. You always get the whole ranked list of up to 100 places back; detail only decides how many of those records come back with hours, photos and popular times attached rather than the ranked core. detail 0–20 costs $0.0150 a call. detail 21–100 costs a flat $0.0750 a call — 21 and 99 are the same price. So detail=20 bills $0.0150 and detail=21 bills $0.0750: 5× for one extra enriched place, with nothing in between. The default is detail=20, inside the cheap band. Do the division:

$0.0150 / 100 places returned         = $0.15 per 1,000 places   (Deep, detail=20 — the default)
$0.0150 /  20 fully-enriched places   = $0.75 per 1,000 places   (Deep, detail=20 — the default)
$0.0750 / 100 fully-enriched places   = $0.75 per 1,000 places   (Deep, detail=100)
$0.0750 /  30 fully-enriched places   = $2.50 per 1,000 places   (Deep, detail=30 — the mid-band trap)
$0.0150 /  20 fully-enriched places   = $0.75 per 1,000 places   (Quick)

Read those five lines together and the shape of the product falls out. Deep at the default gets you a thousand ranked business records carrying name, address, coordinates, rating and — where Google exposes them — phone and website, at $0.15 per 1,000 places returned ($0.105 on the Scale tier). It is also $0.75 per 1,000 fully-enriched records, which is the same rate as Quick and the same rate as Deep at detail=100 — $0.75 per 1,000 enriched at both ends of the range, falling to $0.675 on Growth and $0.525 on Scale.

Here is where the August version of this article overreached, and I would rather correct it in public than quietly restate it. It claimed Serpent landed under "Apify's cheapest per-record actor at $1.00 per 1,000". That $1.00 figure was wrong: scraperlink publishes $0.50 per 1,000 on its free plan and $0.40 on Business, and the fully-enriched rate with popular times is $0.75 to $0.65. So the true comparison on the enriched record is $0.525 (Serpent Scale) against $0.65 (Apify Business, $999/mo), $0.675 (Serpent Growth) against $0.70 (Apify Scale, $199/mo), and $0.75 against $0.75 — an exact tie — at Serpent Default versus Apify's free plan. We win that comparison at volume by about 19%, we win it narrowly in the middle, and at the entry rate we do not win it at all. That is a much smaller claim than the one this article used to make, and it is the one the numbers actually support.

The honest caveat is the middle of the premium band. At detail=30 you pay the full $0.0750 for 30 enriched records — $2.50 per 1,000 fully-enriched, worse than every alternative in the table above except the official Places API at entry volumes. The band is flat, so partial depth is the one place where Serpent is bad value. Buy at 20 or buy at 100. A short response costs more per place either way. New accounts get a shared 10-call free pool usable on every Maps endpoint — enough to evaluate the wire shape before paying anything.

The Cheap End Is Selling a Thinner Record

A 267× price gap between a DataForSEO task and the official Places API's search step — and more than 1,000× against its full search-plus-details pipeline — is not a market inefficiency waiting to be arbitraged. It is a signal that the two are not selling the same object, and the only way to check is to put the field lists side by side.

DataForSEO publishes the field list for a Google Maps SERP item in its own documentation. A row carries title, address, place identifiers, coordinates, phone, URL, rating, rating distribution, category, price level, work hours, a main image, a photo count and a claimed flag. That is a genuinely useful record and for a great many jobs it is the whole job. What it does not carry, and what a Serpent place record does, is popular times, amenities, service options, the full photo array, plus code, timezone and business status. ScrapingDog's Maps response is the same shape: title, address, coordinates, phone, hours, rating, review count, a thumbnail and a photos link — the local-pack row, not the place page behind it.

That is the honest reason the price gap exists, and it cuts both ways. If you are building a lead list and you store name, address, phone, website, rating and hours, DataForSEO at ~$0.006 per 1,000 places is between eighty and a hundred and thirty times cheaper than us and you should use it. I would rather write that sentence than pretend a thin row and a full record are the same purchase. If you need popular times to model footfall, or service options and amenities to filter, or the photo array, or business status to drop closed listings, then those fields are the product and the rate you should be comparing is the fully-enriched one — which is where the field is much narrower: scraperlink at $0.65–$0.75 per 1,000 and Serpent at $0.525–$0.75.

Enrichment is priced as an add-on almost everywhere, which is what makes headline rates misleading. On scraperlink, popular times is +$0.25 per 1,000 on top of a $0.50 base — a 1.5× step. On the official Places API, the search step is $1.60 per 1,000 places and Essentials Place Details adds $5.00 per 1,000 on every one of them — a 4.1× step. Compare on the fully-enriched record, and write down the exact fields you will store before you pick anyone.

Field coverage matters for this math too, because a field you are quoted is not a field you receive. On Serpent, the August 2026 measurement across 97 complete places: business status, opening hours, images, rating, and plus code at 100%; amenities 96%; website 92%; phone 70%. Phone is the gap that matters for lead-gen — roughly a third of places simply do not expose one, on any provider. The provider comparison goes field by field.

How to Decide

Match the provider to the record, not to the headline. I run one of these, so I have written the routing the way I would actually give it — including the four cases where the answer is not us.

• You only need the local-pack row — name, address, phone, website, rating, hours. → DataForSEO, at about $0.006 per 1,000 places ($0.60 per 1,000 tasks, up to 100 results a task). Nothing here comes close, us included. If you would rather pay per request than run a task queue, ScrapingDog at ~$0.05 per 1,000 places on the $40/mo Lite plan is the simpler version of the same trade.
• You need every record fully enriched — popular times, amenities, service options, photos, plus code, timezone, business status. → Serpent Maps Deep at detail=100 or Maps Quick, both $0.75 per 1,000 fully-enriched at Default and $0.525 on Scale, against Apify scraperlink at $0.75 on its free plan and $0.65 on the $999/mo Business plan. Do not stop at a partial depth like detail=30: our premium band is flat, so that is $2.50 per 1,000 fully-enriched and worse than every alternative in the table above except the official API at entry volumes.
• You want a cheap bulk export of ranked places and will enrich only the winners. → Serpent Maps Deep at the default detail=20: $0.15 per 1,000 places returned, $0.105 on Scale.
• You need first-party Google data with a Cloud contract behind it. → The Places API, at about $6.60 per 1,000 businesses with Essentials details or $18.60 with Pro details — and materially less if you are past 5 million calls a month.
• You prefer flat per-record billing with no call-shape maths. → Bright Data at $1.30–$1.50 per 1,000 records, or Outscraper at $3.00 per 1,000 falling to $1.00 above 100,000. Both cost several times the cheap end; you are buying simplicity.
• You need many engines on one meter and one vendor relationship. → SerpApi, at $1.25 per 1,000 places on the $25/mo Starter plan.

At-volume math from the published rates above, Serpent at Default. Your real mix of search / place / reviews calls changes the total. Two things worth noticing. First: on ranked rows only, DataForSEO is 25× cheaper than us and the honest answer is to use them. Second: enriching all 10,000 places inside Deep (detail=100, ~$7.50) is less than half the cost of the list-plus-per-record Place route (~$16.50) — the Place endpoint is for refreshing one high-value record, not for bulk enrichment.

FAQ

What is the cheapest Google Maps API in 2026?

It depends which record you are buying, and no provider wins both units. If you only need the local-pack row of name, address, phone, website, rating and hours, DataForSEO is the cheapest I could source: $0.60 per 1,000 tasks on the standard queue, and one task returns up to 100 results, which is about $0.006 per 1,000 places. ScrapingDog is next, at $0.24 to $1.00 per 1,000 requests with 20 places a request, so roughly $0.012 to $0.05 per 1,000 places. Neither row carries popular times, amenities, service options, the photo array, plus code, timezone or business status. On the fully-enriched record the cheapest rate I could source is Serpent at $0.525 per 1,000 places on the Scale tier, against Apify scraperlink at $0.65 per 1,000 on its $999 a month Business plan and $0.75 on its free and Starter plans. Serpent's Default rate of $0.75 exactly matches that free-plan price. Every figure was read on the vendor's own page on 8 September 2026.

Is the official Google Places API expensive?

At entry volumes, yes. Nearby Search and Text Search bill $32.00 per 1,000 calls on the Pro tier, and one search call returns at most 20 places because maxResultCount is capped at 20, so the search step alone is $1.60 per 1,000 places. Place Details is charged separately, once per place, at $5.00 per 1,000 on Essentials or $17.00 on Pro, so a search-plus-details pipeline runs about $6.60 to $18.60 per 1,000 businesses. Google does publish steep volume steps: above 5 million calls a month the marginal rates fall to $2.40 per 1,000 searches and $0.38 per 1,000 Essentials details, which is about $0.50 per 1,000 businesses. That is a marginal rate at a volume almost nobody reaches, not an entry price.

Why do per-record prices hide the real cost?

Because the headline rate buys the thin row and the useful fields are line items on top. Apify's scraperlink actor charges $0.50 per 1,000 records on its free and Starter plans, then adds $0.25 per 1,000 for popular times and $0.25 per 1,000 review records, so a fully-enriched place is $0.75 rather than $0.50. The official Places API has the same shape at a different scale: $1.60 per 1,000 places for the search step, then another $5.00 per 1,000 for Essentials Place Details on each one, which is roughly four times the base. Compare on the full enriched record and on the exact fields you will store, not on the headline rate.

How were these prices verified?

Every competitor figure in this article was read on that vendor's own live pricing or documentation page on 8 September 2026, and the sources are listed at the foot of the article. Serpent's numbers come from our live billing model. An earlier version of this post carried four wrong competitor figures and a claim about SerpApi credit expiry that their pricing page does not support; all of them are corrected or removed here. Prices change constantly, so treat this as a dated snapshot rather than a standing guarantee.

Is per-call or per-record pricing cheaper?

It depends on how full each call is, and on which records you count. A Serpent Deep call bills one of two flat rates set by the detail parameter: $0.0150 for detail 0 to 20, and $0.0750 for detail 21 to 100. The default $0.0150 call returns up to 100 ranked places, which is $0.15 per 1,000 places returned, or $0.75 per 1,000 if you count only the 20 fully-enriched records. Per-record pricing is more predictable, which is its real advantage, but enrichment is usually a separate line item there too.

How does SerpApi's credit model work?

It is a monthly search subscription rather than pay-as-you-go. Starter is $25 a month for 1,000 searches, which is $25.00 per 1,000 searches, and a Google Maps search returns up to 20 local results, so about $1.25 per 1,000 places. Volume plans bring that down: the largest published Cloud plan works out near $1.96 per 1,000 searches, or roughly $0.098 per 1,000 places, at $106,050 a month. On their pricing page as of 8 September 2026 there is no statement that unused searches expire at month-end. What it does say is that on a downgrade any remaining searches from the previous plan are added to an Extra Credits balance, and that if you cancel you can keep using remaining searches until the renewal date. An earlier version of this article claimed a month-end expiry; I could not source it, so it is gone.