GPU Rental Prices
SearchCtrl K

H200 NVL rental price: from $3.62/hr

The cheapest on-demand H200 NVL is $3.62 an hour at Massed Compute, checked 21 November 2026.

Cheapest on-demand · 143 GB · checked 21 November 2026$3.62/GPU-hourMassed Compute
  • Highest on-demand $3.79
  • 2 offers
  • 2 providers
  • Cheapest unchanged since 7 July
  • Same offers, 90 days: unchanged
  • Source massedcompute.com

H200 NVL offers: 2 configurations from 2 providers

More filters: provider, memory, node size, stock, sort, price unit

2 offers; rates are per GPU and the node total is for the smallest listed node.

2 providers with on-demand offers, fresh prices first, then cheapest; out-of-stock offers last. Tap a provider for its configurations.

GPU / providerBilling$/GPU-hourNode · total per hourRent
H200 NVL · Massed Compute141 GBon-demand$3.621× $3.62Rent →
checked 21 NovDetails
Checked
21 November 2026

Source price ↗ · Report this price

H200 NVL · RunPod143 GBsecure cloud$3.79–Out of stock
checked 21 NovDetails
Checked
21 November 2026

Source price ↗ · Report this price

Fresh prices come first. Within each billing type, prices run lowest first, with unavailable offers last. Commercial relationships never affect ordering. How we collect and rank prices.

Same GPU, provider, billing and price on several hosts or listings: one row, with the count. – means the provider does not state it.

Cite this data

GPU Rental Prices, "H200 NVL rental price: from $3.62/GPU-hour", prices checked 21 November 2026, https://gpurentalprices.com/h200-nvl

Quote it with a link to this page. Downloads of the whole dataset are on Data & API.

On-demand and secure-cloud rates are capacity the provider runs, with no reservation. Spot and community capacity can be interrupted. A marketplace host is one independent owner's server listed on a marketplace. Prices are per GPU per hour, before storage and data transfer. Every row shows the date it was checked.

H200 NVL price history and range

Date range for the H200 NVL chart
  • Cheapest on-demand$3.62(unchanged since 7 July)
  • Provider median$3.79
Daily, 2026-07-06 to 2026-11-21, from our own price checks. The provider median counts each provider's cheapest on-demand rate once and is drawn only on days with three or more providers. Missing days stay gaps.

H200 NVL price range by billing type

Per GPU-hour, latest checked offers. Axis $0 to $4.

On-demand2 offers · 2 providersfrom $3.62to $3.79

CheapestProvider median (each provider's cheapest rate counted once; shown with three or more providers)

Price changes

  • 30 days: unchanged across 2 matching on-demand configurations since 22 October 2026.
  • 90 days: unchanged across 2 matching on-demand configurations since 23 August 2026.

Changes use the geometric mean of price ratios for the same provider, variant, tier, GPU count, instance and region (RunPod's place is where that day's stock sat, so it is not matched). New listings are not counted as price changes. Full ledger on the history hub.

H200 NVL daily figures (139 days)
DateCheapest on-demandProvider median
2026-11-21$3.62–
2026-11-20$3.62–
2026-11-19$3.62–
2026-11-18$3.62–
2026-11-17$3.62–
2026-11-16$3.62–
2026-11-15$3.62–
2026-11-14$3.62–
2026-11-13$3.62–
2026-11-12$3.62–
2026-11-11$3.62–
2026-11-10$3.62–
2026-11-09$3.62–
2026-11-08$3.62–
2026-11-07$3.62–
2026-11-06$3.62–
2026-11-05$3.62–
2026-11-04$3.62–
2026-11-03$3.62–
2026-11-02$3.62–
2026-11-01$3.62–
2026-10-31$3.62–
2026-10-30$3.62–
2026-10-29$3.62–
2026-10-28$3.62–
2026-10-27$3.62–
2026-10-26$3.62–
2026-10-25$3.62–
2026-10-24$3.62–
2026-10-23$3.62–
2026-10-22$3.62–
2026-10-21$3.62–
2026-10-20$3.62–
2026-10-19$3.62–
2026-10-18$3.62–
2026-10-17$3.62–
2026-10-16$3.62–
2026-10-15$3.62–
2026-10-14$3.62–
2026-10-13$3.62–
2026-10-12$3.62–
2026-10-11$3.62–
2026-10-10$3.62–
2026-10-09$3.62–
2026-10-08$3.62–
2026-10-07$3.62–
2026-10-06$3.62–
2026-10-05$3.62–
2026-10-04$3.62–
2026-10-03$3.62–
2026-10-02$3.62–
2026-10-01$3.62–
2026-09-30$3.62–
2026-09-29$3.62–
2026-09-28$3.62–
2026-09-27$3.62–
2026-09-26$3.62–
2026-09-25$3.62–
2026-09-24$3.62–
2026-09-23$3.62–
2026-09-22$3.62–
2026-09-21$3.62–
2026-09-20$3.62–
2026-09-19$3.62–
2026-09-18$3.62–
2026-09-17$3.62–
2026-09-16$3.62–
2026-09-15$3.62–
2026-09-14$3.62–
2026-09-13$3.62–
2026-09-12$3.62–
2026-09-11$3.62–
2026-09-10$3.62–
2026-09-09$3.62–
2026-09-08$3.62–
2026-09-07$3.62–
2026-09-06$3.62–
2026-09-05$3.62–
2026-09-04$3.62–
2026-09-03$3.62–
2026-09-02$3.62–
2026-09-01$3.62–
2026-08-31$3.62–
2026-08-30$3.62–
2026-08-29$3.62–
2026-08-28$3.62–
2026-08-27$3.62–
2026-08-26$3.62–
2026-08-25$3.62–
2026-08-24$3.62–
2026-08-23$3.62–
2026-08-22$3.62–
2026-08-21$3.62–
2026-08-20$3.62–
2026-08-19$3.62–
2026-08-18$3.62–
2026-08-17$3.62–
2026-08-16$3.62–
2026-08-15$3.62–
2026-08-14$3.62–
2026-08-13$3.62–
2026-08-12$3.62–
2026-08-11$3.62–
2026-08-10$3.62–
2026-08-09$3.62–
2026-08-08$3.62–
2026-08-07$3.62–
2026-08-06$3.62–
2026-08-05$3.62–
2026-08-04$3.62–
2026-08-03$3.62–
2026-08-02$3.62–
2026-08-01$3.62–
2026-07-31$3.62–
2026-07-30$3.62–
2026-07-29$3.62–
2026-07-28$3.62–
2026-07-27$3.62–
2026-07-26$3.62–
2026-07-25$3.62$3.79
2026-07-24$3.62$3.79
2026-07-23$3.62$3.79
2026-07-22$3.62$3.79
2026-07-21$3.62$3.79
2026-07-20$3.62$3.79
2026-07-19$3.62$3.79
2026-07-18$3.62$3.79
2026-07-17$3.62$3.79
2026-07-16$3.62$3.79
2026-07-15$3.62$3.79
2026-07-14$3.62$3.79
2026-07-13$3.62$3.79
2026-07-12$3.62$3.79
2026-07-11$3.62$3.79
2026-07-10$3.62$3.79
2026-07-09$3.62$3.79
2026-07-08$3.62$3.79
2026-07-07$3.62$3.79
2026-07-06$3.79–

H200 NVL prices by provider and billing type

Each provider's cheapest rate in each billing type, per GPU-hour. The provider name opens its H200 NVL page with node sizes, fees and its own history.

ProviderOn-demand
Massed Compute$3.62
RunPod$3.79

H200 NVL hardware specs

H200 NVL hardware specifications
ArchitectureHopper
VRAM143 GB
FP16 TFLOPS (dense)835
Memory bandwidth4,800 GB/s
TDP600 W
InterconnectNVLink 900GB/s
Launched2024

Manufacturer datasheet figures for the H200 NVL. Source: www.nvidia.com.

H200 NVL variants

LLM memory estimates for this GPU

Mixtral 8x7B v0.1 at 16-bit weights
112.7GB
Weights 93.4 GB · cache 0.5 GB · headroom 18.8 GB. Capacity line: 143 GB.

These models have an estimated inference footprint within 143 GB at one or more precisions. The scenario uses 4,096 tokens, batch 1, a 16-bit KV cache and 20% runtime headroom. A memory fit does not establish speed or software compatibility.

Model16-bit GB8-bit GB4-bit GB
Qwen3 235B A22B565.2 283.1 142.0 ✓
Qwen3 235B A22B 2507565.2 283.1 142.0 ✓
Mixtral 8x22B v0.1338.6 169.9 85.5 ✓
GLM 4.5 Air266.1 133.5 ✓67.2 ✓
Qwen2.5 72B176.1 88.9 ✓45.2 ✓
DeepSeek R1 Distill Llama 70B170.9 86.3 ✓43.9 ✓
Mixtral 8x7B v0.1112.7 ✓56.7 ✓28.7 ✓
Qwen2.5 32B79.9 ✓40.6 ✓20.9 ✓
Qwen2.5 Coder 32B79.9 ✓40.6 ✓20.9 ✓
DeepSeek R1 Distill Qwen 32B79.9 ✓40.6 ✓20.9 ✓
QwQ 32B79.9 ✓40.6 ✓20.9 ✓
Qwen3 32B79.9 ✓40.6 ✓20.9 ✓
Qwen3 30B A3B73.8 ✓37.1 ✓18.8 ✓
Qwen3 30B A3B 250773.8 ✓37.1 ✓18.8 ✓
Qwen3 Coder 30B A3B73.8 ✓37.1 ✓18.8 ✓
Mistral Small 3.2 24B 250658.4 ✓29.6 ✓15.2 ✓
Mistral Small 24B 250157.4 ✓29.1 ✓14.9 ✓
Devstral Small 250757.4 ✓29.1 ✓14.9 ✓
Qwen2.5 14B36.4 ✓18.7 ✓9.8 ✓
Qwen2.5 Coder 14B36.4 ✓18.7 ✓9.8 ✓
DeepSeek R1 Distill Qwen 14B36.4 ✓18.7 ✓9.8 ✓
Qwen3 14B36.2 ✓18.5 ✓9.7 ✓
phi 436.2 ✓18.6 ✓9.8 ✓
Phi 4 reasoning36.2 ✓18.6 ✓9.8 ✓
Phi 4 reasoning plus36.2 ✓18.6 ✓9.8 ✓
Mistral Nemo 240730.2 ✓15.5 ✓8.2 ✓
Qwen3 8B20.4 ✓10.6 ✓5.6 ✓
DeepSeek R1 Distill Llama 8B19.9 ✓10.3 ✓5.5 ✓
Qwen2.5 7B18.6 ✓9.4 ✓4.9 ✓
Qwen2.5 Coder 7B18.6 ✓9.4 ✓4.9 ✓
DeepSeek R1 Distill Qwen 7B18.6 ✓9.4 ✓4.9 ✓
Mistral 7B v0.318.0 ✓9.3 ✓5.0 ✓
Qwen3 4B10.4 ✓5.6 ✓3.1 ✓
Qwen3 4B 250710.4 ✓5.6 ✓3.1 ✓
Qwen2.5 3B7.6 ✓3.9 ✓2.0 ✓
Qwen2.5 Coder 3B7.6 ✓3.9 ✓2.0 ✓
Qwen3 1.7B5.4 ✓3.0 ✓1.8 ✓
DeepSeek R1 Distill Qwen 1.5B4.4 ✓2.3 ✓1.2 ✓
Qwen2.5 1.5B3.8 ✓2.0 ✓1.1 ✓
Qwen2.5 Coder 1.5B3.8 ✓2.0 ✓1.1 ✓
Qwen3 0.6B2.4 ✓1.5 ✓1.0 ✓
Qwen2.5 0.5B1.2 ✓0.7 ✓0.4 ✓
Qwen2.5 Coder 0.5B1.2 ✓0.7 ✓0.4 ✓

Models that need more than one H200 NVL

Compare model memory requirements or change context and memory assumptions on a model's page.

When to choose the H200 NVL, and what to pick instead

Start from memory: your model, its context and your batch must fit in 143 GB (the model-fit table above estimates it). Then the H200 NVL is the right card unless one of these fits your job better.

Rent or buy the H200 NVL

Confirmed H200 NVL purchase listings start at $30,576 (H200 NVL 141GB, SabrePC, checked 21 November 2026, out of stock). That sum rents about 8,446 GPU-hours at the cheapest on-demand rate of $3.62 (checked 21 November 2026), about 4.1 years at 40 hours a week, before electricity, hosting and resale.

Compare H200 NVL purchase prices and quote availability, then enter your usage and electricity costs in the rent-versus-buy calculator. Purchase listings have their own variants, stock status and update dates. Compare two GPUs.

H200 NVL questions

How much does it cost to rent an H200 NVL?
The cheapest H200 NVL rental we have is $3.62 per GPU-hour on Massed Compute (on-demand), checked 21 November 2026.
How much VRAM does an H200 NVL have?
143 GB.
Where is the H200 NVL cheapest right now?
Massed Compute has the cheapest checked on-demand rate on 21 November 2026 at $3.62 per GPU-hour.

GPUs near the H200 NVL

Related

All 140 GB+ GPUs · GPUs under $2/hr · Price movers this week · Massed Compute review

Compare your offers

Saved prices keep their original dates. Confirm the configuration, terms and availability at the source.

Select up to three offers to compare their rates and terms.

Type a GPU, a provider, a model or a question.