Free BigQuery Cost Calculator

Estimate your monthly Google BigQuery spend in seconds. Model both pricing models - on-demand (per-TB scanned) and capacity Editions (Standard / Enterprise / Enterprise Plus slot-hours) - side by side with storage tiering and streaming ingest included. Perfect for budget planning, on-demand-vs-Editions break-even analysis, and multi-cloud cost comparison against Snowflake and Databricks.

How BigQuery pricing works

BigQuery splits cost into three completely independent dimensions: compute (how queries run), storage (how tables are kept on disk), and streaming ingest (how new rows arrive). Compute is almost always the dominant cost. Storage is usually cheap. Streaming is niche and often free at small volumes.

On-demand pricing ($6.25 per TB scanned)

The default billing model. You pay $6.25 for every terabyte of uncompressed logical data your queries scan. The first 1 TB of scan per project per month is free. No cluster to provision, no reservation to manage - just pay per query. Best for small and intermittent workloads, dev / staging projects, and any team with unpredictable usage. Downside: a single bad query can cost hundreds of dollars if it scans a full table without partitioning.

Capacity (Editions) pricing

You buy slot-hours in advance at one of three tiers:

• Standard: $0.04/slot-hour. No BigQuery ML, no materialized views. Entry-level SQL.
• Enterprise: $0.06/slot-hour. Full feature set (ML, materialized views, column-level security). Most common tier.
• Enterprise Plus: $0.10/slot-hour. Adds autoscaling slots (0 to max), cross-region + cross-cloud queries, CMEK, assured workloads.

Storage tiers

• Active storage - $0.02/GB/month - any table or partition modified in the last 90 days.
• Long-term storage - $0.01/GB/month (50% off) - automatic on tables/partitions not modified for 90+ days. No query penalty, totally automatic.
• Physical storage (optional, opt-in per dataset) - pay for actual compressed bytes instead of logical. Typically ~50% cheaper for well-compressed datasets.

When to switch from on-demand to capacity Editions

Break-even example: You are running 450 TB/month on on-demand. That costs (450 - 1) * $6.25 = $2,806/month. A Standard Edition reservation of 100 slots running 24/7 costs 100 * 730 * $0.04 = $2,920/month - about the same. But if your workload is bursty (big overnight ETL, idle during the day), Enterprise Plus with autoscale (50-150 avg slots) often comes in at ~$2,200/month, saving 20-40%.

General rule: if your on-demand spend is above $3,000/month and your query patterns are predictable, move to Editions. If spend is below $1,000/month or highly intermittent, stay on-demand.

How to reduce BigQuery cost on on-demand

Bytes scanned is the only cost lever on on-demand. Apply these in order of ROI:

1. Partition every fact table. Use DATE partitioning on the event timestamp for almost all analytics data. Filtering on the partition column prunes months of data in a single WHERE clause - often 90%+ bytes saved.
2. Cluster on common WHERE/GROUP BY columns. Up to 4 columns. Clustering prunes blocks within a partition.
3. Never SELECT *. BigQuery is columnar - unused columns are not scanned. Specifying columns can cut scan 50-90% on wide tables.
4. Use approximate aggregations. APPROX_COUNT_DISTINCT is ~100x cheaper than COUNT(DISTINCT) on large tables when 2% error is acceptable.
5. Materialize hot joins. A materialized view or scheduled query that pre-joins and pre-aggregates the two highest-traffic tables can eliminate 80% of dashboard scan cost.

BigQuery vs Snowflake vs Databricks

For predictable medium workloads the three platforms are within ~20% of each other on list price. Real differences come from usage patterns. Use the Snowflake Cost Calculator and Databricks Cost Calculator for apples-to-apples comparisons at your actual TB and workload shape.

Related tools

Use the JSON to SQL DDL Generator to scaffold BigQuery CREATE TABLE statements from sample data. Use the SQL Formatter to clean up dialect-specific SQL. For multi-cloud comparisons see Snowflake vs BigQuery.

← Back to Home