NVIDIA · Hopper · 2022
NVIDIA H100. Rental by the hour, on demand.
From $3.15/hr per GPU, pay as you go with no commitment. The Hopper workhorse for production training and inference: 80 GB of HBM3, up to 8 GPUs in one node.
H100 on demand
- Availability
- Available now · 6 regions
- By region and size
- $3.15 to $8.57/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 machine
80 GB of HBM3. At up to 3.4 TB/s.
- Memory
- 80 GB
- HBM3
- Bandwidth
- up to 3.4 TB/s
- Memory to compute
- FP16 tensor
- up to 989
- TFLOPS
- Per node
- 8 GPUs
- up to 700 W each
Price across clouds
NVIDIA H100, cloud by cloud. Sorted by price.
On-demand rates per GPU-hour from each cloud's public price page on 8 Nov 2026, for their H100 SXM: RunPod's Secure Cloud, and Lambda's lowest per-GPU rate across its cluster sizes. Runcrate's row is today's live starting rate.
Workloads
What teams run on the H100. Each one opens the launch form, sized.
Launch it
Ask your agent for an H100. 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.
Start an H100 on Runcrate called my-gpu, copy train.py to it and run it.
create_instance { name: "my-gpu", gpu: "H100" }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 H100 compares. Against its closest alternatives.
Benchmarks
The H100, measured. Real runs, dated, with what they cost.
- LLM inference
Qwen2.5-7B on vLLM, three GPUs (run 2026-10-26); FP8 vs bf16 on an H100 (run 2026-10-27)
- A100 vs H100 vs H200
One script on six GPUs (run 2026-10-25)
- Speech to text
Whisper large-v3 audio hours per dollar (run 2026-10-28)
- Text embeddings
BGE embedding throughput per dollar (run 2026-10-29)
- Training and fine-tuning
torch.compile vs eager for training (run 2026-10-29)
- Image generation
SDXL images per dollar on four GPUs (run 2026-10-27)
Alternatives
Alternatives to the H100. Same account, different price and performance.
| GPU | Memory · FP16 tensor | Stands out | Price gap | From |
|---|---|---|---|---|
| NVIDIA H100This GPU | 80 GB HBM3 · up to 989 TFLOPS | $3.15/hr | ||
| NVIDIA H200 | 141 GB HBM3e · 989 TFLOPS | +61 GB memory | +36% | $4.28/hr |
| NVIDIA GH200 | 96 GB HBM3 · 990 TFLOPS | Not in stock right now | ||
| NVIDIA H200 NVL | 141 GB HBM3e · 835 TFLOPS | +61 GB memory | +48% | $4.67/hr |
| AMD MI300X | 192 GB HBM3 · 1307.4 TFLOPS | +112 GB memory | +54% | $4.85/hr |
Full specs
NVIDIA H100 specifications.
Memory
- Memory
- 80 GB
- Type
- HBM3
- Bandwidth
- up to 3.4 TB/s
Compute
- FP32
- up to 67 TFLOPS
- FP16 tensor
- up to 989 TFLOPS
- INT8
- up to 1979 TOPS
- Tensor cores
- up to 528
- CUDA cores
- up to 16,896
Power and form
- TDP
- up to 700 W
- Architecture
- Hopper
- Released
- 2022
Node
- GPUs per node
- 8
- Maker
- NVIDIA
Questions
Renting the H100. What buyers ask first.
Other GPUs
Browse every GPU by architecture.
NVIDIA Blackwell
NVIDIA Ampere
NVIDIA Ada Lovelace
NVIDIA Volta