BigQuery Cost Calculator
Start with how much data you have, or with the bill you already pay. Google's prices for every region, checked October 2026.
How much data will you keep in BigQuery?
All the tables together, before compression. Rough is fine. In rows, we count about 1 KiB a row: a table of 20 to 40 columns of numbers, dates and short text.
How Querylab.io lowers this bill
Querylab.io is a BigQuery editor. On $30 a month, its cost features typically take off $12.
- Limits per query and per dayA query over your limit is stopped before it runs.−$1.19
- Warnings while you typeCostly patterns are flagged before the query runs.−$1.13
- Tablesample modeWrite and test on a sample, run on all of the data once.−$2.15
- Table and column analysisShows which tables and columns your queries pay to read.−$1.93
- Query cost insightsThe price and hints for every query in your history.−$0.87
- Storage billing modelShows which datasets cost less billed by compressed size.−$4.68
- Storage analysisWhat each table costs to keep, so you can drop what nobody needs.−$0.19
Typical results for teams using each feature, not a promise for yours. Each saving applies to what is left after the ones above it.
How BigQuery pricing works
A BigQuery bill has two main parts. Queries are billed for the data they process, or for the slots you rent if you use capacity pricing. Storage is billed for the data you keep, by the GiB per month.
With on-demand pricing, the cost of a query depends on how many bytes it reads, not on how long it runs or how many rows it returns. BigQuery stores each column separately, so a query reads only the columns it uses. On a partitioned table, a filter on the partition column also limits which partitions it reads. More in how queries are billed.
Storage is billed by uncompressed size unless you switch a dataset to physical billing, which charges for compressed bytes at about twice the price per GiB. Data usually compresses several times over, so for most datasets physical billing costs less.
Prices in US multi-region
| Item | Price (USD) |
|---|---|
| On-demand queries | $6.25 per TiB processed, first 1 TiB a month free |
| Active storage, logical | $0.02 per GiB a month |
| Long-term storage, logical | $0.01 per GiB a month |
| Active storage, physical | $0.04 per GiB a month |
| Long-term storage, physical | $0.02 per GiB a month |
| Standard edition slots | $0.04 per slot-hour |
| Enterprise edition slots | $0.06 per slot-hour, $0.048 with 1 year, $0.036 with 3 years |
| Enterprise Plus edition slots | $0.1 per slot-hour, $0.08 with 1 year, $0.06 with 3 years |
List prices from Google's Cloud Billing Catalog, fetched 2026-10-30, in USD at Google's own rate. What you pay is in the currency of your billing account. Change the region at the top of the calculator to see another. The first 10 GiB of storage each month is free. Streaming inserts, BI Engine and data transfer are billed separately.
Questions
How much does a 1 TiB query cost in BigQuery?
With on-demand pricing, BigQuery charges $6.25 for each TiB a query reads in the US multi-region. The first 1 TiB each month is free, so a query that reads 1 TiB is free the first time in a month and costs $6.25 after that.
Are other regions cheaper?
Yes, prices differ by region. Queries cost $6.25 per TiB in EU multi-region, US multi-region, Iowa and 7 more, and up to $11.25 in São Paulo. Storage ranges from $0.02 to $0.037 per GiB a month. US multi-region charges $6.25 per TiB and $0.02 per GiB. Queries run where the data is stored, so moving to a cheaper region means copying the data there.
Is BigQuery free?
Every account gets 1 TiB of query reading and 10 GiB of storage free each month. Small projects often stay inside that and pay nothing.
How do I know how much a query reads?
BigQuery shows it before you run a query, above the editor: "This query will process 1.2 GB when run." That number is what an on-demand query is billed for. Querylab.io shows the same estimate in dollars while you type.
Does LIMIT make a query cheaper?
No. BigQuery reads the columns a query uses before the LIMIT applies, so the bill is the same. The exception is a clustered table, where BigQuery can stop reading early. Selecting fewer columns and filtering on the partition column are what make a query read less.
Do repeated queries cost money?
Running the same query again within about 24 hours is usually served from the query cache, and cached results are not billed. The cache is skipped when the data has changed or the query uses functions like CURRENT_TIMESTAMP.
When is capacity pricing cheaper than on-demand?
Capacity pricing charges for slots by the hour, whatever the queries read. It tends to win when many queries read a lot of data every day, usually once on-demand queries pass about $1,500 a month. Below that, on-demand usually costs less.
How can I lower a BigQuery bill?
Make queries read less: select only the columns you need, filter on the partition column, and try a query on a sample before running it on everything. Check what a query will process before you run it. For storage, compare logical and physical billing for each dataset, and drop or expire tables nobody uses. Querylab.io shows the cost of each query as you type and can stop the ones over your limit.
Further reading: free tier and sandbox, on-demand vs capacity pricing, the query cache, controlling costs.
