NVIDIA H200 rental pricing

The cheapest NVIDIA H200 in stock right now is $2.09 per GPU-hour on-demand from Beam for the 1× H200 SXM configuration. The median on-demand rate across 27 providers is $4.20 per GPU-hour.

Running one NVIDIA H200 around the clock for a month (730 hours) costs about $1,526 at the cheapest in-stock rate of $2.09 per hour. At the median on-demand rate it is about $3,064 a month. Reserved terms usually come in below that; ask the provider for a quote.

Beam has the cheapest NVIDIA H200 in stock right now at $2.09 per GPU-hour on-demand for the 1× H200 SXM configuration. Rates change often, so check the provider table for today's figure before you commit.

We have 5 weeks of price history for the NVIDIA H200, not yet enough like-for-like weeks to call a trend: the set of providers we track is still growing, and a median over more sellers is not a price move. The chart above fills in as our daily checks accumulate.

On-demand is pay-as-you-go with no commitment. Spot is spare capacity sold at a discount that the provider can reclaim at short notice. Reserved is a committed term, usually months, in exchange for a lower rate. Community is capacity from individual hosts on a marketplace, cheaper and less uniform. This week's median rates for the NVIDIA H200 are on-demand $4.11 (28 providers), spot $2.61 (4 providers), reserved $3.63 (5 providers), community $3.59 (4 providers), per GPU-hour.

We track 32 providers with 61 live prices for the NVIDIA H200 across 2 variants. Sold-out listings stay in the table so you can see who normally carries it, but never set the cheapest figure.

The SXM is the cheapest NVIDIA H200 variant right now at $2.09 per GPU-hour. Cheapest in-stock rate per variant: SXM from $2.09, NVL from $3.29. The variants differ in memory, bandwidth and interconnect, so the cheapest card is not always the cheapest job.

With 8-bit weights and a single sequence of context, one NVIDIA H200 holds gpt-oss-20b (20.9B), Gemma 3 27B (27.4B), Qwen3-32B (32.8B), Llama 3.3 70B Instruct (70.6B), gpt-oss-120b (117B). See the model fit table for the full ladder.

At its cheapest in-stock on-demand rate the NVIDIA H200 delivers dense FP8 compute at $1.06 per PFLOP-hour, which ranks 9 of 30 priced GPUs we cover on that basis. Lower is better: it is the rental price divided by peak FP8 throughput.

Prices are read daily from each provider's public price list or API and recorded whenever they change. Listings are shown per GPU-hour on-demand unless marked otherwise, and the chart covers the last 30 days. The newest listing was checked just now.