CoreWeave instance types and prices

One row per instance type, showing the lowest price compute.info collected for each billing term and currency, with the region it comes from. Billing terms and currencies are never mixed or converted. This is the scope compute.info collects, not CoreWeave's complete catalog: a missing size or region is not a claim that CoreWeave does not sell it.

CoreWeave: lowest collected price per instance type and billing term, in native currency.
Instance typeResourcesOn-demand (lowest region)Other billing terms
NVIDIA GH200Family: gpu2 regions listed72 vCPU · 480 GiB RAM1× NVIDIA GH200 · 96 GiB/cardCPU allocation not specified · Architecture not specifiedStorage not specified
USD 6.50 / hr≈$4,745.00 / mo · 730-hour estimate, not a capEurope price table (no data center stated) · lowest of 2 regionsTax treatment not specified · source
None collected
NVIDIA L40Family: gpu2 regions listed128 vCPU · 1024 GiB RAM8× NVIDIA L40 · 48 GiB/cardCPU allocation not specified · Architecture not specifiedStorage not specified
USD 10.00 / hr≈$7,300.00 / mo · 730-hour estimate, not a capEurope price table (no data center stated) · lowest of 2 regionsTax treatment not specified · source
None collected
NVIDIA L40SFamily: gpu2 regions listed128 vCPU · 1024 GiB RAM8× NVIDIA L40S · 48 GiB/cardCPU allocation not specified · Architecture not specifiedStorage not specified
USD 18.00 / hr≈$13,140.00 / mo · 730-hour estimate, not a capEurope price table (no data center stated) · lowest of 2 regionsTax treatment not specified · source
None collected
NVIDIA RTX PRO 6000 Blackwell Server Edition (High Memory)Family: gpu2 regions listed128 vCPU · 1024 GiB RAM8× NVIDIA RTX PRO 6000 Blackwell Server Edition · 96 GiB/cardCPU allocation not specified · Architecture not specifiedStorage not specified
USD 20.00 / hr≈$14,600.00 / mo · 730-hour estimate, not a capEurope price table (no data center stated) · lowest of 2 regionsTax treatment not specified · source
None collected
NVIDIA A100Family: gpu2 regions listed128 vCPU · 2048 GiB RAM8× NVIDIA A100 · 80 GiB/cardCPU allocation not specified · Architecture not specifiedStorage not specified
USD 21.60 / hr≈$15,768.00 / mo · 730-hour estimate, not a capEurope price table (no data center stated) · lowest of 2 regionsTax treatment not specified · source
None collected
NVIDIA HGX B200Family: gpu2 regions listed128 vCPU · 2048 GiB RAM8× NVIDIA B200 · 180 GiB/cardCPU allocation not specified · Architecture not specifiedStorage not specified
USD 68.80 / hr≈$50,224.00 / mo · 730-hour estimate, not a capEurope price table (no data center stated) · lowest of 2 regionsTax treatment not specified · source
None collected
NVIDIA HGX H100Family: gpu2 regions listed128 vCPU · 2048 GiB RAM8× NVIDIA H100 · 80 GiB/cardCPU allocation not specified · Architecture not specifiedStorage not specified
USD 49.24 / hr≈$35,945.20 / mo · 730-hour estimate, not a capEurope price table (no data center stated) · lowest of 2 regionsTax treatment not specified · source
None collected
NVIDIA HGX H200Family: gpu2 regions listed128 vCPU · 2048 GiB RAM8× NVIDIA H200 · 141 GiB/cardCPU allocation not specified · Architecture not specifiedStorage not specified
USD 50.44 / hr≈$36,821.20 / mo · 730-hour estimate, not a capEurope price table (no data center stated) · lowest of 2 regionsTax treatment not specified · source
None collected
NVIDIA GB200 NVL72Family: gpu2 regions listed144 vCPU · 960 GiB RAM4× NVIDIA GB200 · 186 GiB/cardCPU allocation not specified · Architecture not specifiedStorage not specified
USD 42.00 / hr≈$30,660.00 / mo · 730-hour estimate, not a capEurope price table (no data center stated) · lowest of 2 regionsTax treatment not specified · source
None collected

CoreWeave billing and source notes

Official CoreWeave pricing page "On-demand GPU instances" tables, one per region (North America and Europe): on-demand prices per instance per hour, shown with "$" and recorded as USD. The page does not state a tax basis. Monthly amounts are 730-hour estimates. Products listed as "Contact sales" have no published price and are not shown; spot and inference-platform prices are out of scope.

Provider-wide egress terms are not specified in the collected data.

View vendor pricing at CoreWeave ↗ · CoreWeave product page ↗

Data source for these prices

This is the original collection source for auditing our data. It may contain raw API/JSON data or require authentication; it is not a buying page.

Original collection source

Spot prices can be interrupted, and commitments and reservations bind you to a term. Neither is comparable to on-demand. A vendor monthly cap is not hourly price multiplied by 730. Read the methodology before purchasing, or filter the complete VM catalog by region, hardware and billing term across every provider.