GPU compute markets
Providers
Boost Run
GPU cloud. 6 GPU models listed right now, via Shadeform.
GPUs listed
US dollars per GPU-hour, against each GPU's median across every provider. Largest memory first.
| GPU | VRAM | Tier | Price | Market median | Against it | Supply | Updated |
|---|---|---|---|---|---|---|---|
| B200sold out | 180 GB | on-demand | $3.74/h | $7.99/h | -53% | Out of stock | 53 min ago |
| H200 SXMsold out | 141 GB | on-demand | $4.25/h | $4.52/h | -6% | Out of stock | 53 min ago |
| RTX PRO 6000sold out | 96 GB | on-demand | $0.99/h | $1.97/h | -50% | Out of stock | 53 min ago |
| L40Ssold out | 48 GB | on-demand | $0.69/h | $1.26/h | -45% | Out of stock | 53 min ago |
| RTX 6000 Adasold out | 48 GB | on-demand | $0.60/h | $0.80/h | -25% | Out of stock | 53 min ago |
| RTX 4090 | 24 GB | on-demand | $0.49/h | $0.40/h | +22% | In stock | 53 min ago |
Machines
6 whole machines, cheapest per GPU first. A machine sold in several regions of one area shows once, at its cheapest.
| GPU | Tier | GPUs | vCPUs | RAM | Region | Per GPU | Machine |
|---|---|---|---|---|---|---|---|
| RTX 4090 | on-demand | 8 | 64 | 512 GB | US, Raleigh, NC | $0.49/h | $3.92/h |
| RTX 6000 Ada | on-demand | 8 | 128 | 512 GB | US, Seattle, WA · none free | $0.60/h | $4.80/h |
| L40S | on-demand | 8 | 64 | 1536 GB | US, Dallas, TX +1 more · none free | $0.69/h | $5.52/h |
| RTX PRO 6000 | on-demand | 8 | 128 | 1536 GB | US, Charlotte, NC · none free | $0.99/h | $7.92/h |
| B200 | on-demand | 8 | 96 | 2048 GB | US, Dallas, TX +1 more · none free | $3.74/h | $29.92/h |
| H200 SXM | on-demand | 8 | 96 | 2048 GB | US, Raleigh, NC +1 more · none free | $4.25/h | $34.00/h |