How much does it cost to run a 12 gig query in BigQuery?
Updated: 20 November 2026Request a BigQuery review quoteFor an Iowa (us-central1) on-demand query scanning 12 GiB of billable data, the calculation is about $0.0732 before the free allowance. Confirm what « 12 gig » measures and which billing model applies.

Calculate 12 GB and 12 GiB separately
At the verified Iowa (us-central1) on-demand rate of $6.25 per TiB, a 12 GiB billable scan costs about $0.0732 before allowances. If « 12 gig » means 12 decimal GB, the equivalent calculation is about $0.0682. These are query-processing calculations for the stated rate, not a quote for an analytics platform or its storage.
The official BigQuery price sheet, checked on 20 November 2026, lists this USD rate for Iowa (us-central1) and the first 1 TiB of on-demand query data processed per month per account as free. Verify your location, currency and agreement. If the applicable free allowance still covers this scan, its incremental on-demand query-processing charge can be zero.
| Meaning of 12 gig | Conversion to TiB | Processing at $6.25 per TiB |
|---|---|---|
| 12 GiB: 12 × 2^30 bytes | 12 / 1024 = 0.01171875 TiB | 0.01171875 × $6.25 = $0.0732421875 |
| 12 GB: 12,000,000,000 bytes | 12,000,000,000 / 2^40 ≈ 0.01091394 TiB | Approximately $0.06821210 |
| Eligible usage fully covered by the remaining free allowance | No paid query bytes for this scan | Zero incremental query-processing charge |
Use the job's billed-byte quantity for the final reconciliation. Google rounds processing upward to the next MB, with a 10 MB minimum per referenced table and per query. The decimal-GB row is a unit-conversion estimate before those billing adjustments, rather than an exact job charge. A dataset described as 12 GB does not establish that a particular query bills 12 GB. Selected columns and the data actually read matter. The BigQuery pricing and cost-control guide covers the wider compute and storage boundary.
Check what the query will actually read
Google's query cost-estimation guidance documents the query validator and dry runs for estimating bytes processed. Compare the estimate with the intended columns, filters and table layout before running unfamiliar SQL. For clustered tables, the pre-execution estimate is an upper bound, so retain the completed job's billed bytes for reconciliation.
- Confirm that the job uses on-demand pricing and identify the billing account and location.
- Clarify whether the size means decimal GB, GiB, table size, result size or billed scan.
- Use the query validator or a dry run to inspect the estimated processing volume.
- Review available allowance and the applicable rate or contract.
- Compare the completed job metadata with the estimate and retain the evidence.
Result size and scan size answer different questions. Google's cost-control guidance says LIMIT does not reduce bytes read for non-clustered tables. For clustered tables, it can reduce scanned bytes when scanning stops after enough blocks return the requested rows. The pricing documentation excludes queries retrieving cached results from on-demand query charges. Check whether a job actually used the cache before treating a repeated execution as another paid scan.
For model operations, use the separate BigQuery ML pricing boundary. For application requests that trigger analytics, combine the query review with the Cloud Run cost review while keeping each billed service separate.
Turn a small scan into a workload estimate
The unit calculation is only the first input. Ask how often the query runs, whether its scanned data grows and which executions use cached results. Review scheduled reports, retries and interactive use separately. Do not multiply one observed job by a guessed monthly execution count and present the result as measured spend.
| Workload question | Evidence | Review outcome |
|---|---|---|
| How frequently does the job execute? | Job history and scheduler configuration | A supported execution baseline |
| Does every run read the same volume? | Billed bytes across representative runs | An explicit volume assumption |
| Which charges sit outside this scan? | Storage, streaming and other service billing | A complete analytics cost boundary |
| Who can approve a restrictive setting? | Analytics owner and change process | An accepted guardrail and exception route |
A useful review connects a billed job to a business output. Request query identifiers, reproducible calculations and a record of approved changes. Compare candidates on their ability to explain scan behaviour and preserve required reporting. The FinOps proposal checklist helps separate diagnostic work from implementation.
For organization-wide ownership and billing reconciliation, start with the GCP cost optimization scope. If the requirement is to reject expensive on-demand queries, use the BigQuery cost-control matrix to evaluate maximum bytes billed and daily quotas with the analytics owner.