Snowflake Warehouse Sizes and Credits per Hour (2026)

Snowflake warehouses come in ten sizes, from X-Small to 6X-Large. An X-Small uses 1 credit per hour and each size up doubles that, so a 6X-Large uses 512. Gen2 warehouses, now the default for new standard warehouses, use 1.35 times as many credits per hour on AWS and GCP and 1.25 times on Azure. This guide gives the full tables in credits and in dollars, explains how Snowflake bills each size, and shows how to check and change a warehouse’s size. We checked every fact and figure in this guide against Snowflake’s documentation.

This article is part of our Snowflake pricing guide.

What warehouse sizes does Snowflake offer?

Snowflake offers ten standard warehouse sizes. Each size has twice the compute of the one below it and also uses twice the credits per hour. However, Gen2 warehouses come in X-Small to 4X-Large only.

Source: Snowflake’s Service Consumption Table, Tables 1(a) and 1(b), effective 15 October 2026. Also, 5X-Large and 6X-Large may not be available in every region.

How much does each warehouse size cost per hour?

Multiply the credits per hour by your price per credit. For example, at Snowflake’s on-demand list price for Enterprise Edition in AWS US East, $3.00 per credit:

On Standard Edition ($2.00) it is a third lower. However, a capacity contract replaces these list prices with the rate in your order form. See how much a Snowflake credit costs.

How does Snowflake bill warehouse time?

Snowflake bills a warehouse per second while it runs, with a 60-second minimum each time it starts or resumes. It keeps billing while idle until it suspends, and a suspended warehouse costs nothing.

• Idle time counts. A new warehouse suspends after 600 seconds of inactivity by default, so a short burst of queries can be followed by up to 10 minutes of paid idle time.
• Each resume costs at least 60 seconds. A warehouse that wakes for a 5-second query still pays for a full minute.
• Resizing takes effect for new work. Resizing a running warehouse does not affect queries already running; they finish on the old size, and the new size serves queued and new queries. Billing for the added compute starts the moment you resize.

Does a bigger warehouse make queries faster?

Often, but not always. Snowflake’s documentation says query performance generally scales with warehouse size, particularly for larger, more complex queries, but that larger is not necessarily faster for small, basic queries, and that a bigger warehouse does not always load data faster. If a query runs in half the time on a warehouse twice the size, it costs the same, because the bigger warehouse uses twice the credits for half as long. If it does not speed up, for example because it is small or waits on something else, the bigger size just costs more.

As a result, size is a trade-off between speed and waste, not a fixed choice. Our guides on vertical scaling and warehouses that never got resized back cover how to test it.

Gen1 vs Gen2: which generation should you use?

Gen2 standard warehouses are Snowflake’s newer generation of compute and are now the default when you create a standard warehouse, except in the few regions where Gen2 is not offered. They use more credits per hour (1.35 times on AWS and GCP, 1.25 times on Azure), so they only save money when queries finish fast enough to make up the difference.

• Gen2 is available from X-Small to 4X-Large.
• Snowflake does not offer Gen2 in AWS EU (Zurich), AWS Africa (Cape Town), GCP Middle East Central2 (Dammam) or the Azure US Government Virginia regions.
• You can convert an existing warehouse with ALTER WAREHOUSE … SET GENERATION = ‘2’. If it is running, queries already in progress finish on Gen1 and you pay for both sets of compute while they do.
• New Gen2 warehouses get Query Acceleration turned on by default on Enterprise and above; converting a Gen1 warehouse does not turn it on.

Our Gen1 vs Gen2 comparison shows how to measure whether Gen2 pays off for your workloads.

What about multi-cluster warehouses?

A multi-cluster warehouse runs several clusters of the same size at once to handle more concurrent queries. Each running cluster bills at the full rate for its size, so a Large warehouse with 3 clusters running uses 3 x 8 = 24 credits per hour on Gen1. Multi-cluster warehouses also need Enterprise Edition or higher.

Three settings decide what a multi-cluster warehouse costs:

• MIN_CLUSTER_COUNT and MAX_CLUSTER_COUNT set how many clusters can run. When they are equal, every cluster runs whenever the warehouse runs.
• SCALING_POLICY = STANDARD (the default) starts extra clusters quickly to avoid queuing.
• SCALING_POLICY = ECONOMY keeps running clusters fully loaded and starts a new one only when there is enough work to keep it busy for at least 6 minutes. As a result, it queues more and spends less.

Snowflake’s guidance is that resizing is not the tool for concurrency: more users running queries at once calls for more clusters, while bigger, slower queries call for a bigger size.

What do Snowpark-optimized and other warehouse types cost?

Snowpark-optimized warehouses give each node more memory, for machine learning and other memory-heavy Snowpark work. The default memory setting starts at Medium and uses more credits per hour than a standard warehouse of the same size.

Two newer types bill differently:

• Interactive warehouses use fewer credits per hour (0.6 for an X-Small, up to 76.8 for a 4X-Large) but have a 60-minute minimum each time they start or resume.
• Adaptive Warehouses have no fixed hourly rate: credits scale with the work they do. They need Enterprise Edition or higher.

How do you check and change a warehouse’s size?

To see each warehouse’s size and suspend setting:

SHOW WAREHOUSES;

A warehouse created with CREATE WAREHOUSE is X-Small unless you set a size. To create, convert and resize:

-- Create a Gen2 Small warehouse that suspends after 60 seconds idle
CREATE WAREHOUSE reporting_wh
  GENERATION = '2'
  WAREHOUSE_SIZE = SMALL
  AUTO_SUSPEND = 60;

-- Convert an existing warehouse to Gen2
ALTER WAREHOUSE etl_wh SET GENERATION = '2';

-- Change a warehouse's size
ALTER WAREHOUSE reporting_wh SET WAREHOUSE_SIZE = 'MEDIUM';

-- Suspend it after 60 seconds of inactivity instead of the default 600
ALTER WAREHOUSE reporting_wh SET AUTO_SUSPEND = 60;

To see which warehouses use the most credits, so you know where size matters most:

-- Credits by warehouse, last 30 days
SELECT warehouse_name,
       SUM(credits_used) AS credits_used
FROM snowflake.account_usage.warehouse_metering_history
WHERE start_time >= DATEADD(day, -30, CURRENT_TIMESTAMP())
GROUP BY 1
ORDER BY 2 DESC;

Which warehouse size should you choose?

Start with the smallest size that meets your performance target, then measure. Move up a size only if queries get meaningfully faster; move down if they barely slow. Signs that a warehouse is too small include queries waiting in a queue and queries spilling to disk. Signs that it is too large include a warehouse that is mostly idle while running, or queries that run no faster than they did one size down.

Workloads also change. For example, a warehouse sized for month-end reporting is often too big the rest of the month. Our guide to automating Snowflake warehouse optimization covers how to keep sizes matched to the work as it changes.

How Seemore helps with warehouse sizing

Seemore’s SmartPulse turns each warehouse’s real usage pattern into an hourly plan: it downsizes through quiet hours, scales back up for recurring peaks, and recommends Gen1 or Gen2 alongside size, so the extra cost of Gen2 is only taken on when it pays for itself. Auto-Shutdown suspends each warehouse the moment it goes idle instead of waiting out a fixed timeout, without touching a running query.