GPU Cloud Pricing · Hopper
The inference-tuned H100 variant: 94 GB of HBM3 and higher bandwidth than the SXM part, purpose-built for serving LLMs in PCIe servers.
GPU.ai rates are the live fleet floor — the cheapest provider with stock, passed through at cost. Competitor rates are public on-demand list prices.
| Provider | $ / GPU / hr |
|---|---|
| GPU.ai — Secure Cloud | $3.19 |
| GPU.ai — Community Cloud | $2.59 |
| CoreWeave | Not available |
| AWS | Not available |
| Azure | Not available |
Available regions: Asia Pacific · Europe West · US Central · US East · US West
GPU.ai prices are read live from our aggregated fleet of 12+ GPU clouds and normalized to a per-GPU hourly rate (a multi-GPU node's price is divided by its GPU count). We charge the provider's raw rate with zero markup and bill per second. Competitor figures are each provider's public on-demand list rate for the closest equivalent SKU — never spot, reserved, or committed-use discounts. Stock counts reflect launchable inventory, not theoretical capacity. Last updated 2026-08-17; this page refreshes hourly.
| Architecture | Hopper |
| VRAM | 94 GB HBM3 |
| Memory bandwidth | 3,900 GB/s |
| FP16 tensor (dense) | 835 TFLOPS |
| CUDA cores | — |
| Power (TDP) | 400 W |
| Interconnect | PCIe 5.0 + NVLink bridge |
| Released | 2023 |
Minimum H100 NVL count needed to hold popular open models in VRAM, at FP16 and 4-bit quantization.
| Model | Params | FP16 | 4-bit |
|---|---|---|---|
| Llama 3.1 8B | 8B | 1× H100 NVL | 1× H100 NVL |
| Mistral 7B | 7B | 1× H100 NVL | 1× H100 NVL |
| Gemma 2 27B | 27B | 1× H100 NVL | 1× H100 NVL |
| Qwen2.5 32B | 32B | 1× H100 NVL | 1× H100 NVL |
| Llama 3.1 70B | 70B | 2× H100 NVL | 1× H100 NVL |
| Qwen2.5 72B | 72B | 2× H100 NVL | 1× H100 NVL |
| Mixtral 8x22B | 141B MoE | 4× H100 NVL | 2× H100 NVL |
| Llama 3.1 405B | 405B | Doesn't fit | 4× H100 NVL |
| FLUX.1 dev | 12B | 1× H100 NVL | 1× H100 NVL |
| SDXL | 3.5B | 1× H100 NVL | 1× H100 NVL |
| Whisper large-v3 | 1.5B | 1× H100 NVL | 1× H100 NVL |
Weights-only estimates with a 20% margin for KV cache and activations at moderate context. Long contexts, large batches, and training need substantially more memory.
H100 NVL cloud instances on GPU.ai currently start at $2.59 per GPU per hour, billed per second. That is the live floor price across every provider in our fleet — we route your launch to the cheapest provider with stock and add zero markup.
The NVIDIA H100 NVL has 94 GB of HBM3 memory with up to 3,900 GB/s of bandwidth. GPU.ai reports the minimum VRAM actually delivered across our fleet, so the figure you see at launch is the figure you get.
GPU.ai aggregates a dozen-plus GPU clouds and passes through each provider's raw price with no markup, so the same silicon is typically far cheaper than hyperscaler list rates. The comparison table on this page shows the current public on-demand list prices side by side.
Billing is per second, starting when your instance is SSH-ready and stopping the moment you terminate. There are no minimum commitments, no reservation fees, and no egress surprises — you pay the listed hourly rate pro-rated to the second.