Skip to content

NVIDIA · Hopper · 2023

NVIDIA H200. Rental by the hour, on demand.

From $5.13/hr per GPU on demand, or prepay an 8-GPU machine at $5.01/GPU-hr for 1 month, $5.01/GPU-hr for 3 months.

See also H200 NVL pricing: the PCIe card, its GPUs joined by NVLink bridges in groups of four.

H200 on demand

from$5.13/hr per GPU
Availability
Available now · 6 regions
By region and size
$5.13 to $10.67/hr
Prepaid
1 mo $5.01/hr3 mo $5.01/hr6 mo $4.72/hr12 mo $4.57/hr
Billing
By the second on most GPUs, a few by the minute
8+ GPUs or a month
Launch any size yourself, or prepay a machine for 1 to 12 months at a discount. Custom quotes on request.

The card

141 GB of HBM3e. At 4.8 TB/s.

Memory
141 GB
HBM3e
Bandwidth
4.8 TB/s
Memory to compute
FP16 tensor
989
TFLOPS
Per node
8 GPUs
NVLink, 700 W each

Price across clouds

NVIDIA H200, cloud by cloud. Sorted by price.

On-demand rates per GPU-hour from each cloud's public price page on 8 Nov 2026: RunPod's Secure Cloud, and Lambda's lowest per-GPU rate across its cluster sizes. Runcrate's row is today's live starting rate.

Runcrate$5.13/hr
RunPod$5.29/hr

Workloads

What teams run on the H200. Each one opens the launch form, sized.

Launch it

Ask your agent for an H200. Or read the call it makes.

Claude Code or Codex runs it for you: it starts the instance, copies your files over and runs the job. The SDKs make the same request.

Connect your agent first

Add Runcrate's MCP server to yourself: run curl -fsSL "https://runcrate.ai/mcp.md" and follow it.

Add to CursorInstall in VS Code
Works inClaude CodeCodexOpenCodeClineRoo CodeKilo CodeCursor

Start an H200 on Runcrate called my-gpu, copy train.py to it and run it.

MCP callstext
create_instance { name: "my-gpu", gpu: "H200" }
instance_status { id: "my-gpu" }
file_upload { id: "my-gpu", content: "<train.py>", path: "/root/workspace/train.py" }
ssh_execute { id: "my-gpu", command: "nohup python /root/workspace/train.py > train.log 2>&1 &" }

Spec sheet

How the H200 compares. Against its closest alternatives.

FP16 tensor

NVIDIA H200989 TFLOPS
NVIDIA H100up to 989 TFLOPS
NVIDIA GH200990 TFLOPS
NVIDIA H200 NVL835 TFLOPS

Memory

NVIDIA H200141 GB
NVIDIA H10080 GB
NVIDIA GH20096 GB
NVIDIA H200 NVL141 GB

Bandwidth

NVIDIA H2004.8 TB/s
NVIDIA H100up to 3.4 TB/s
NVIDIA GH200up to 4.0 TB/s
NVIDIA H200 NVL4.8 TB/s

Alternatives

Alternatives to the H200. Same account, different price and performance.

GPUMemory · FP16 tensorStands outPrice gapFrom
NVIDIA H200This GPU141 GB HBM3e · 989 TFLOPS$5.13/hr
NVIDIA H10080 GB HBM3 · up to 989 TFLOPS-61 GB memory-24%$3.91/hr
NVIDIA GH20096 GB HBM3 · 990 TFLOPS-45 GB memoryNot in stock right now
NVIDIA H200 NVL141 GB HBM3e · 835 TFLOPS-9%$4.67/hr
AMD MI300X192 GB HBM3 · 1307.4 TFLOPS+51 GB memory-5%$4.85/hr

Full specs

NVIDIA H200 specifications.

Memory

Memory
141 GB
Type
HBM3e
Bandwidth
4.8 TB/s

Compute

FP32
67 TFLOPS
FP16 tensor
989 TFLOPS
INT8
1979 TOPS
CUDA cores
16,896

Power and form

TDP
700 W
Form factor
SXM
Architecture
Hopper
Released
2023

Node

GPUs per node
8
Interconnect
NVLink
Maker
NVIDIA

Questions

Renting the H200. What buyers ask first.

Other GPUs

Browse every GPU by architecture.