← back

H100 PCIe

80 GB VRAM
vCPU26RAM221 GB
1x2x
from$1.60/hr

H100 NVL

94 GB VRAM
vCPU64RAM126 GB
1x2x4x
from$1.78/hr

H100 SXM

80 GB VRAM
vCPU24RAM283 GB
1x2x4x8x
from$1.82/hr

H200 NVL

143 GB VRAM
vCPU32RAM181 GB
1x2x4x
from$2.70/hr

H200 SXM

141 GB VRAM
vCPU28RAM252 GB
1x2x4x8x
from$3.05/hr

B200

180 GB VRAM
vCPU24RAM250 GB
1x2x4x8x
from$4.20/hr

A40

48 GB VRAM
vCPU1+RAM5 GB+
1x2x4x
from$0.06/hr

A10

24 GB VRAM
vCPU16RAM63 GB
1x2x4x
from$0.18/hr

A100 40GB PCIe

40 GB VRAM
vCPU48RAM252 GB
1x2x4x8x
from$0.65/hr

A100 40GB SXM

40 GB VRAM
vCPU8RAM63 GB
1x2x4x8x
from$0.90/hr

A100 80GB PCIe

80 GB VRAM
vCPU18RAM215 GB
1x2x
from$1.25/hr

A100 80GB SXM

80 GB VRAM
vCPU16RAM117 GB
1x2x4x8x
from$1.35/hr

GH200

96 GB VRAM
vCPU64RAM432 GB
1x
from$2.05/hr

purchase flow

1. chatSend workload, duration, GPU/CPU target, region preference, and your public SSH key.
2. quoteI confirm availability, final hourly/day rate, estimated start time, and usage limit.
3. payPay via QRIS after the slot is confirmed. No slot is reserved before confirmation.
4. sshAfter payment clears, you get SSH host, username, port, and teardown time.

Better workflow: keep ordering chat-based first. Static checkout is risky for compute because stock, quota, region, and price move fast.

For bigger jobs, use partial upfront + timed extension. For small jobs, full upfront via QRIS is simplest.

ready?
Ask for a live quote and payment QR.

request qris

notes