GPU cloud · head to head · checked 8 August 2026
CoreWeave vs RunPod
6 GPU instances from the two, side by side: what each plan includes, what it costs, and where the price comes from.
CoreWeave
from $18/hr
L40S (8x) · 3 plans listed
Best price per GB of VRAM: $0.034/GB VRAM per hour (A100 80GB (8x))
RunPod
from $0.34/hr
RTX 4090 · 3 plans listed
Best price per GB of VRAM: $0.014/GB VRAM per hour (RTX 4090)
Which is cheaper, CoreWeave or RunPod?
RunPod's cheapest listed plan is RTX 4090 at $0.34/hr; CoreWeave's is L40S (8x) at $18/hr.
Which gives more for the money, CoreWeave or RunPod?
On price per GB of VRAM, RunPod does best with RTX 4090 at $0.014/GB VRAM per hour; CoreWeave's best is A100 80GB (8x) at $0.034/GB VRAM per hour.
How many GPU instances does each offer here?
We list 3 for CoreWeave and 3 for RunPod, each checked against the provider's own pricing page.
CoreWeave
Everything from CoreWeaveL40S (8x)
- GPU
- NVIDIA L40S
- VRAM
- 384 GB
- Tier
- On-demand
- Region
- US
A100 80GB (8x)
- GPU
- NVIDIA A100 SXM
- VRAM
- 640 GB
- Tier
- On-demand
- Region
- US
HGX H100 (8x)
- GPU
- NVIDIA H100 SXM
- VRAM
- 640 GB
- Tier
- On-demand
- Region
- US
RunPod
Everything from RunPodRTX 4090
- GPU
- NVIDIA RTX 4090
- VRAM
- 24 GB
- Tier
- Community Cloud (spot, can be interrupted)
- Region
- 30+ regions
A100 PCIe
- GPU
- NVIDIA A100 PCIe
- VRAM
- 80 GB
- Tier
- Community Cloud (spot, can be interrupted)
- Region
- 30+ regions
H100 PCIe
- GPU
- NVIDIA H100 PCIe
- VRAM
- 80 GB
- Tier
- Community Cloud (spot, can be interrupted)
- Region
- 30+ regions
CoreWeave against the rest
RunPod against the rest