review galore

Price check · Nov 6, 2026

CoreWeave Cloud pricing

No free tier or per-seat plan: budget by instance-hour, workload, and whether you use spot or on-demand compute.

4.0/10Value score

We may earn a commission if you buy through links on this page. It never changes a score or a ranking. How we make money

Short answer

Paid plans start at $3.27/mo.

CoreWeave Cloud has no free plan or published free trial. Pricing is per instance-hour, not per seat, so a one-to-five-person team pays for the compute it runs. The lowest listed rate is $3.27 per hour for an Intel CPU spot instance; the single-GPU NVIDIA GH200 is $6.50 per hour on demand. Three or five users do not automatically mean three or five instances. Monthly and annual totals are not listed, so estimate usage hours and confirm other charges with CoreWeave.

01Every plan, line by line

PlanBilled annuallyBilled monthlyBuilt for
NVIDIA GB300 NVL72—QuoteLarge-scale AI training
NVIDIA GB200 NVL72—QuoteLarge-scale AI training
NVIDIA HGX B300—QuoteLarge-scale AI training
NVIDIA HGX B200—QuoteLarge-scale AI training
NVIDIA RTX PRO 6000 Blackwell Server Edition (High Memory)—QuoteGPU workloads needing high memory
NVIDIA RTX PRO 6000 Blackwell Server Edition (Standard Memory)—QuoteGPU workloads needing standard memory
NVIDIA GH200—QuoteSingle-GPU workloads
NVIDIA HGX H100—QuoteAI training and inference
NVIDIA L40—QuoteGPU rendering and inference
NVIDIA HGX H200—QuoteAI training and inference
NVIDIA L40S—QuoteGPU rendering and inference
NVIDIA A100—QuoteAI training and inference
AMD Genoa (9274F), High Performance—QuoteHigh-performance CPU workloads
AMD Genoa (9454), General Purpose—QuoteGeneral-purpose CPU workloads
AMD Genoa (9454), General Purpose (High Storage)—QuoteCPU workloads needing more storage
AMD Genoa (9654), High Core—QuoteHigh-core CPU workloads
AMD Turin (9655P), General Purpose (High Memory)—QuoteCPU workloads needing high memory
AMD Turin (9655P), General Purpose (Standard Memory)—QuoteGeneral-purpose CPU workloads
AMD Turin (9655P), General Purpose (High Storage)—QuoteCPU workloads needing more storage
Intel Emerald Rapids (8562Y+), General Purpose (High Memory)—QuoteCPU workloads needing high memory
Intel Emerald Rapids (8562Y+), General Purpose (Standard Memory)—QuoteGeneral-purpose CPU workloads

Add-ons and overages

  • Inference GPU rates are listed for North America: GB200 $10.50/hr; HGX B200 $8.60/hr; RTX PRO 6000 High Memory $2.50/hr; GH200 $6.50/hr; H100 $6.16/hr; L40 $1.25/hr; H200 $6.31/hr; L40S $2.25/hr; A100 $2.70/hr. Contact sales for other listed inference rates.

02What the plan grid doesn't tell you

Storage and networking

The platform includes AI storage and networking, but their prices are not listed here. Ask for rates based on your expected data and cluster traffic.

Inference rates

Some North America inference GPU rates are listed separately from compute rates. Confirm the rate and configuration for the model you plan to serve.

Spot versus on-demand

Some spot instances are cheaper than on-demand, while others have no listed spot rate. Confirm which rate applies to your region and workload.

Unpublished billing terms

The pricing page lists hourly rates, not monthly or annual totals. Ask about billing commitments and any other charges before forecasting spend.

03Which plan should you buy?

If you are…GetWhy
Solo CPU workloadIntel Emerald Rapids (8562Y+), General Purpose (High Memory)The listed spot rate is $3.27 per hour; on-demand is $5.31 per hour.
Small GPU workloadNVIDIA GH200One GPU with 96 GB VRAM is listed at $6.50 per hour on demand; spot is unavailable.
Multi-GPU trainingNVIDIA HGX H100Eight GPUs with 80 GB VRAM; listed at $49.24 per hour on demand or $19.71 spot in North America.

04Is CoreWeave Cloud worth the money?

CoreWeave Cloud can make sense for a small team if GPU infrastructure is essential and the team can match instance-hours to real workloads. The platform offers many GPU choices, but there is no free plan, no per-seat pricing, and no published monthly or annual total. The $3.27-per-hour entry point is a CPU spot instance, not a free way to test GPU workloads. Get a quote for storage, networking, and any configuration without a listed rate before committing.

Cheaper CoreWeave Cloud alternatives →

CoreWeave Cloud pricing FAQ

What is the cheapest CoreWeave Cloud price?

The lowest listed compute rate is $3.27 per instance-hour for an Intel Emerald Rapids (8562Y+), General Purpose (High Memory) spot instance. Its on-demand rate is $5.31 per hour. That is a CPU instance, not a GPU plan.

How much does CoreWeave Cloud cost for three people?

There is no published three-person price because CoreWeave lists instance-hour rates rather than seat prices. At the cheapest listed spot rate, three separate instances would total $9.81 per hour. That assumes one instance per person; shared access and seat limits are not listed.

Does CoreWeave Cloud offer monthly pricing?

The pricing page lists hourly on-demand and spot rates, not monthly or annual billing. Your total depends on the instance type and hours used. Ask CoreWeave whether it offers billing commitments or quotes that fit your planned workload.

Is there a free CoreWeave Cloud trial?

CoreWeave Cloud does not have a free plan, and the vendor does not publish a free trial. Compute is billed at listed instance-hour rates. Ask CoreWeave about any evaluation access or account requirements before planning a test.