Nvidia H200

Nvidia H200

Extends the H100 with doubled memory for large-model inference and training.

Compare vs other GPUs →
Aggregating historical prices...

Weekly median price per GPU per hour. Nvidia H200 price history (JSON).

At a glance

Hardware

Architecture
Hopper
Memory per GPU
141 GB HBM3e
Memory bandwidth
4,800 GB/s
Release date
Q4 2023

Market Updated 2 minutes ago

Cheapest
$1.99 / GPU / hr
on-demand · not verified in stock
Median price
$4.44 / GPU / hr
on-demand
90-day trend
Flat
median on-demand
Coverage
41 providers
69 of 223 configs in stock

H200 Pricing and Availability

41 providers

How this list works

The two views

By provider shows one card per company, placed where its best offer ranked. By configuration lists every offer.

Order

What you can rent comes first: in stock, then waitlist, then not reported, then out of stock. Priced offers rank ahead of quote-only, and on-demand ahead of other billing types.

Within a group, five factors set the order:

  1. Location: datacenter proximity, blended with provider HQ.
  2. Price: hourly price, per GPU and in total.
  3. Billing type: reserved ahead of spot, spot ahead of quote-only.
  4. Specs: more VRAM, vCPUs and RAM.
  5. Provider diversity: a provider's repeat rows rank slightly lower, so one company can't take all the top spots.

Search

We match the provider, its country, GPU model, form factor, billing type, availability and instance name. Matching is partial and case-insensitive. Everyday words work too, so "interruptible" finds spot and "sold out" finds out of stock.

Transparency and funding

  • Ads and sponsors: Paid placements sit at the top of the list and are always labeled as sponsored content. Sponsorship never influences the organic ranking itself.
  • Affiliates: Affiliate links are marked. We may earn a commission if you click them, but commissions never affect the order.
  • Prices: Shown in USD, converted at daily reference rates where a provider publishes in another currency. A month means 720 hours. How we estimate costs.
Vast.ai logo

Vast.ai

Our sponsor

United States of America flag USA 47 configs 1x-8x

On-Demand from $3.34 Reserved from $4.18 Spot from $0.40
From $3.34 / GPU / hr On-Demand NVL Visit website
Runpod logo

Runpod

In stock

United States of America flag USA 8 configs 1x-8x

On-Demand from $3.59 Reserved on request
From $3.59 / GPU / hr On-Demand SXM Visit website
Theta EdgeCloud logo

Theta EdgeCloud

In stock

United States of America flag USA 3 configs 1x-4x

On-Demand from $3.69
From $3.69 / GPU / hr On-Demand Visit website
Koyeb logo

Koyeb

In stock

France flag France 4 configs 1x-8x

On-Demand from $3.00
From $3.00 / GPU / hr On-Demand Visit website
GPU.ai logo

GPU.ai

In stock

United Arab Emirates flag UAE 6 configs 1x-8x

On-Demand from $3.34
From $3.34 / GPU / hr On-Demand NVL Visit website
Spheron logo

Spheron

In stock

Singapore flag Singapore 2 configs 1x

On-Demand from $5.92 Spot from $3.36
From $5.92 / GPU / hr On-Demand SXM Visit website
Runcrate logo

Runcrate

In stock

United States of America flag USA 1 config 8x

On-Demand from $4.40
From $4.40 / GPU / hr On-Demand SXM Visit website
Sesterce logo

Sesterce

In stock

France flag France 1 config 8x

On-Demand from $4.40
From $4.40 / GPU / hr On-Demand Visit website
GPUaaS logo

GPUaaS

In stock

United States of America flag USA 1 config 8x

Custom on request
Pricing On request Custom SXM Visit website
Packet·ai logo

Packet·ai

In stock

United States of America flag USA 1 config 8x

Custom on request
Pricing On request Custom NVL Visit website
Together AI logo

United States of America flag USA 4 configs 1x

On-Demand from $2.99 Reserved from $3.99
From $2.99 / GPU / hr On-Demand SXM Visit website
Beam logo

United States of America flag USA 2 configs 1x-8x

On-Demand from $1.99 Custom on request
From $1.99 / GPU / hr On-Demand SXM Visit website
Cerebrium logo

United States of America flag USA 4 configs 1x-8x

On-Demand from $4.20
From $4.20 / GPU / hr On-Demand Visit website
Fal.ai logo

United States of America flag USA 2 configs 1x

On-Demand from $4.50 Custom from $2.10
From $4.50 / GPU / hr On-Demand Visit website
Geodd logo

United States of America flag USA 5 configs 1x-8x

On-Demand from $3.39
From $3.39 / GPU / hr On-Demand NVL Visit website
Replicate logo

United States of America flag USA 4 configs 1x-8x

On-Demand from $5.49
From $5.49 / GPU / hr On-Demand Visit website
Enverge logo

United States of America flag USA 1 config 1x

On-Demand from $5.50
From $5.50 / GPU / hr On-Demand Visit website
DigitalOcean logo

United States of America flag USA 3 configs 1x-8x

On-Demand from $4.47 Reserved from $3.40
From $4.47 / GPU / hr On-Demand SXM Visit website
Hugging Face logo

United States of America flag USA 4 configs 1x-8x

On-Demand from $5.00
From $5.00 / GPU / hr On-Demand Visit website
Civo logo

United Kingdom flag UK 10 configs 1x-8x

On-Demand from $3.49 Reserved from $2.99
From $3.49 / GPU / hr On-Demand SXM Visit website
Hyperstack logo

United Kingdom flag UK 2 configs 1x

On-Demand from $3.99 Reserved from $2.79
From $3.99 / GPU / hr On-Demand SXM Visit website
Nebius logo

Netherlands flag Netherlands 2 configs 1x

On-Demand from $4.50 Spot from $2.45
From $4.50 / GPU / hr On-Demand SXM Visit website
Crusoe logo

United States of America flag USA 3 configs 8x

On-Demand from $4.29 Reserved on request Spot on request
From $4.29 / GPU / hr On-Demand SXM Visit website
Canopy Wave logo

United States of America flag USA 3 configs 2x-8x

On-Demand from $4.00
From $4.00 / GPU / hr On-Demand SXM Visit website
Gcore logo

Luxembourg flag Luxembourg 1 config 8x

On-Demand from $3.17
From $3.17 / GPU / hr On-Demand SXM Visit website
CoreWeave logo

United States of America flag USA 2 configs 8x

On-Demand from $6.30 Spot from $2.62
From $6.30 / GPU / hr On-Demand SXM Visit website
Amazon Web Services logo

AWS

United States of America flag USA 1 config 8x

On-Demand from $7.91
From $7.91 / GPU / hr On-Demand Visit website
OVHcloud logo

OVH

France flag France 1 config 8x

On-Demand from $6.20
From $6.20 / GPU / hr On-Demand Visit website
Oracle Cloud logo

United States of America flag USA 1 config 8x

On-Demand from $10.00
From $10.00 / GPU / hr On-Demand Visit website
Google Cloud logo

United States of America flag USA 4 configs 8x

On-Demand from $10.62 Reserved from $4.66 Spot from $6.37
From $10.62 / GPU / hr On-Demand SXM Visit website
Microsoft Azure logo

United States of America flag USA 4 configs 8x

On-Demand from $13.78 Reserved from $6.86 Spot from $13.78
From $13.78 / GPU / hr On-Demand Visit website
AceCloud logo

India flag India 24 configs 1x-4x

On-Demand from $5.02 Reserved from $3.66
From $5.02 / GPU / hr On-Demand NVL Visit website
Contabo logo

Germany flag Germany 3 configs 1x-8x

Reserved from $2.97
From $2.97 / GPU / hr Reserved (1mo) Visit website
Cirrascale logo

United States of America flag USA 4 configs 8x

Reserved from $3.68
From $3.68 / GPU / hr Reserved (12mo) Visit website
Leaseweb logo

Netherlands flag Netherlands 3 configs 1x-2x

Reserved from $2.29
From $2.29 / GPU / hr Reserved (1mo) Visit website
Oblivus logo

United States of America flag USA 5 configs 8x

On-Demand from $5.50 (sold out) Reserved from $4.47
From $4.47 / GPU / hr Reserved (36mo) SXM Visit website
CUDO Compute logo

United Kingdom flag UK 1 config 1x

Custom on request
Pricing On request Custom SXM Visit website
Green AI Cloud logo

Sweden flag Sweden 1 config 8x

Custom on request
Pricing On request Custom SXM Visit website
Lyceum logo

Germany flag Germany 9 configs 1x-8x

On-Demand from $4.29 (sold out) Reserved on request Spot from $1.60 (sold out)
Pricing On request Reservation Visit website
Massed Compute logo

Massed Compute

Out of stock

United States of America flag USA 8 configs 1x-8x

On-Demand from $3.62 Spot from $3.44
From $3.62 / GPU / hr On-Demand NVL Visit website
Verda logo

Verda

Out of stock

Finland flag Finland 28 configs 1x-8x

On-Demand from $4.00 Reserved from $3.00 Spot from $2.00
From $4.00 / GPU / hr On-Demand SXM Visit website

Heads up: A provider's own page may quote a different figure, for example on tax, region or a promotion. A monthly-only plan shows a derived hourly rate. Verify before provisioning. More on how we price.

What size AI models can the H200 run?

One H200 has 141 GB of VRAM. In practice, that's enough memory for roughly 220B parameters at 4-bit or 58B at 16-bit, assuming a 32K context. Below are some open-weight LLMs, with the estimated memory and GPUs each one needs.

Model Parameters 4-bitINT4 / FP4 8-bitFP8 / INT8 16-bitFP16 / BF16
Alibaba Cloud logo Qwen3.8-27B
27B
1 GPU 18 GB VRAM · $3,197/mo
1 GPU 30 GB VRAM · $3,197/mo
1 GPU 58 GB VRAM · $3,197/mo
Google Cloud logo Gemma 4 31B
30.7B
1 GPU 22 GB VRAM · $3,197/mo
1 GPU 35 GB VRAM · $3,197/mo
1 GPU 69 GB VRAM · $3,197/mo
OpenAI logo GPT-OSS-120B
117B 5.1B active
1 GPU 67 GB VRAM · $3,197/mo
1 GPU 120 GB VRAM · $3,197/mo
2 GPUs 237 GB VRAM · $6,394/mo
DeepSeek logo DeepSeek V4 Flash
284B 13B active
2 GPUs 160 GB VRAM · $6,394/mo
4 GPUs 287 GB VRAM · $12,787/mo
8 GPUs 573 GB VRAM · $25,574/mo
Z.AI logo GLM-5.3-Flash
320B 18B active
2 GPUs 178 GB VRAM · $6,394/mo
4 GPUs 322 GB VRAM · $12,787/mo
8 GPUs 642 GB VRAM · $25,574/mo
MiniMax logo MiniMax-M3
428B 23B active
2 GPUs 239 GB VRAM · $6,394/mo
4 GPUs 432 GB VRAM · $12,787/mo
8 GPUs 862 GB VRAM · $25,574/mo
Moonshot AI logo Kimi K3
2.8T 104B active
13 GPUs 1,544 GB VRAM
23 GPUs 2,804 GB VRAM
45 GPUs 5,605 GB VRAM

Estimates based on the median on-demand rate. Memory is weights plus FP8 KV cache at 32K context per request (FP16/BF16 in the 16-bit column). GPU counts assume 90% of advertised VRAM is usable. No guarantee of runtime support, usable performance, or that a matching quantized build exists. How we estimate costs.

Technical Specifications

Nvidia H200 · Per GPU

Compute · dense
FP8 1,979 TFLOPS 3,958 with sparsity
FP16 / BF16 989.5 TFLOPS 1,979 with sparsity
INT8 1,979 TOPS 3,958 with sparsity
FP32 67 TFLOPS
FP64 34 TFLOPS
Precision support FP8FP16BF16TF32FP32FP64INT8
Memory
Capacity 141 GB HBM3e
Bandwidth 4,800 GB/s
ECC Yes
Silicon
Architecture Hopper
Process TSMC 4N
Transistors 80 billion
Shader cores 16,896 CUDA cores
Matrix cores 528 Tensor cores
Compute units 132 SMs
Fabric and host
GPU interconnect NVLink 900 GB/s
Host interface PCIe 5.0 x16
Power
Board power 700 W
Platform
Partitioning MIG, up to 7 instances

Source: official Nvidia H200 datasheet.

Frequently Asked Questions

How much does the H200 cost per hour?

As of September 5, 2026, the median on-demand price is $4.44 per GPU per hour across 30 providers with a priced on-demand config.

Billing type Configs Median / GPU / hr Cheapest in stock
On-demand 101 $4.44 $1.99
Reserved 80 $4.19 $2.29 (1 mo)
Spot 36 $2.99 $0.40 (Vast.ai)
Custom contract 6 $2.10

No median for custom-contract: we only show this when at least 3 providers list the GPU on that billing type. Custom-contract pricing is negotiated per deal and usually not published.

Who has the cheapest H200?

As of September 5, 2026, the lowest listed on-demand price for the H200 is $1.99 per GPU per hour from Beam, though we haven't verified current stock. Cheapest verified in stock: on-demand $3.00 from Koyeb and spot $0.40 from Vast.ai.

How much does the H200 cost per month?

As of September 5, 2026, at 720 hours per month, one H200 costs an estimated $3,197 at the median on-demand price. Cheapest verified in stock: $2,160 per month on-demand, $3,010 reserved, $288 spot.

Where can I rent an H200?

GetDeploying currently tracks H200 configs from 41 providers. The cheapest verified in-stock on-demand configs come from Koyeb, Vast.ai, GPU.ai and Runpod. See the full price comparison above for every provider and config.

Are H200 prices going up or down?

As of September 5, 2026, the median on-demand price has been flat over the past 90 days, at about $4.44 per GPU per hour, though it is about 23% above where it was a year ago.

How many H200 GPUs do I need?

One H200 runs models up to roughly 220B parameters at 4-bit quantization or 58B at 16-bit, assuming a 32K context. Larger models run across multiple GPUs: the model table above shows the estimated memory and GPU count for popular open-weight LLMs.

Why choose the H200?

141GB HBM3e doubles the memory of the H100 while maintaining NVLink 900 GB/s. Fits 70B+ models without sharding. Best for memory-bound workloads like long-context inference and large batch training.

When is the H200 not a good fit?

Limited availability and premium pricing. If your model fits in 80GB or the workload isn't memory-bottlenecked, the H100 offers better price-per-GPU-hour.

Alternatives to Nvidia H200

Last updated