Loading prices…
GB200 rental prices
NVIDIA · Blackwell · 186 GB HBM3e · Superchip · 2024
3 providers · 5 configurations · prices as of Nov 9, 2026, 21:02 UTC
Lowest on-demand
$10.5per GPU·h · ≈ $7,665/month
at CoreWeaveMedian on-demand
$16.0/GPU·h
across 3 providers
Lowest spot
—
Reserved from $11.9
FP16/BF16 cost
$4.20/PFLOP·h
#57 of 69 GPUs for value
Providers
3
5 configurations
Price history
Daily GB200 prices per GPU-hour. The shaded period is reconstructed from public archives: the version history of SkyPilot's price catalogs, Internet Archive copies of providers' pricing pages and the open gpu-price-feed project (fewer providers than live data). After that, PetaPrice's own tracking.
Lowest on-demandMedian on-demand
Price by provider
Cheapest On-demand GB200 offer at each provider, per GPU-hour. Shorter and greener is cheaper.
CoreWeave🇺🇸$10.5best
Oracle Cloud🇺🇸$16.0+52%
Azure🇺🇸$27.02.6×
Thin vertical line = median across providers ($16.0).
Every provider we track
Pricing
Size
Host
Provider HQ
3 providers · 5 configurations
Sort
- CoreWeaveOn-demand🇺🇸 HQ in United Statesvia CoreWeave pricing pagefrom · On-demandCheapest$10.5/GPU·hCheapest
- 4× GPU144 vCPU · 960 GB RAM$10.5 /GPU·h$42.0 /hOn-demandNorth AmericaRent ↗
- Oracle CloudOn-demand🇺🇸 HQ in United Statesvia Oracle price list APIfrom · On-demand+52%$16.0/GPU·h+52%
- 4× GPU—$16.0 /GPU·h$64.0 /hOn-demandlist price, all regionsRent ↗
- AzureOn-demandReserved🇺🇸 HQ in United Statesvia Azure Retail APIfrom · On-demand2.6×$27.0/GPU·h2.6×
What can the GB200 run?
Memory for the weights plus one 32K-token sequence of context, how many GPUs hold it, and what that replica costs at the cheapest in-stock on-demand rate.
| Model | Params | Memory | GPUs | $/hr | $/mo |
|---|---|---|---|---|---|
gpt-oss-20b OpenAI · 20.9B | 20.9B · 3.6B active | 20 GiB | 1 | $10.50 at CoreWeave | $7,665 |
Gemma 3 27B Google · 27.4B | 27.4B | 28 GiB | 1 | $10.50 at CoreWeave | $7,665 |
Qwen3-32B Alibaba · 32.8B | 32.8B | 39 GiB | 1 | $10.50 at CoreWeave | $7,665 |
Llama 3.3 70B Instruct Meta · 70.6B | 70.6B | 76 GiB | 1 | $10.50 at CoreWeave | $7,665 |
gpt-oss-120b OpenAI · 117B | 117B · 5.1B active | 110 GiB | 1 | $10.50 at CoreWeave | $7,665 |
Qwen3-235B-A22B Alibaba · 235B | 235B · 22B active | 225 GiB | 2 | $21.00 at CoreWeave | $15,330 |
DeepSeek-R1 DeepSeek · 684B | 684B · 37B active | 640 GiB | 4 | $42.00 at CoreWeave | $30,660 |
Kimi K3 Moonshot AI · 2.78T | 2.78T · 104B active | 2.53 TiB | 16 | $168.00 at CoreWeave | $122,640 |
Assumes 90% of each GPU's memory is usable and rounds up to 1, 2, 4, 8… GPUs for tensor parallelism (up to 64). $/hr is the GPU count times the cheapest in-stock on-demand rate per GPU; check that the provider rents instances that large. Serving many requests at once needs more memory for context.
Alternatives to the GB200
GPUs with similar compute and memory, and their lowest on-demand price right now.
- B200 SXM180 GB · 2.25 PF FP16$5.43/h$2.41/PFLOP·h
- GB300288 GB · 2.5 PF FP16$13.6/h$5.43/PFLOP·h
Instinct MI355X288 GB · 2.52 PF FP16$8.60/h$3.42/PFLOP·h- Instinct MI350X288 GB · 2.31 PF FP16$6.16/h$2.67/PFLOP·h
- B300 SXM288 GB · 2.25 PF FP16$5.59/h$2.48/PFLOP·h
- Instinct MI300X192 GB · 1.31 PF FP16$2.39/h$1.83/PFLOP·h
GB200 pricing questions
How much does it cost to rent an GB200?
On-demand GB200 prices start at $10.5 per GPU-hour (CoreWeave). The median across 3 providers is $16.0, and the most expensive is $27.0. Running one GPU around the clock costs about $7,665 per month at the lowest price.
How much compute do you get per dollar with the GB200?
At $10.5 per hour and 2.5 PF of dense FP16/BF16, the GB200 costs $4.20 per PFLOP-hour. That ranks 57 of 69 GPUs with on-demand prices on PetaPrice. Figures are datasheet peaks; real utilization is typically 30–60%.
What are the GB200's key specs?
186 GB HBM3e with 8,000 GB/s of memory bandwidth, Blackwell architecture (2024), Superchip form factor, 1200 W. Dense peak: 2.5 PF FP16 / BF16, 5 PF FP8, 10 PF FP4, 5 PF INT8, 1.25 PF TF32, 0.08 PF FP32, 0.045 PF FP64.
How current are these prices?
Prices were collected on Nov 9, 2026, 21:02 UTC from provider APIs and public catalogs and refresh about every 2 hours. They are indicative list prices; always confirm on the provider's website before booking.