Nvidia B200

Nvidia B200

High-end Blackwell GPU for large-scale AI training and inference.

Last update 5 minutes ago
Compare vs other GPUs →
Aggregating historical prices...

Key Specifications

Architecture
Blackwell
Memory
Up to 192GB HBM3e
Memory Bandwidth
8,000 GB/s
Release date
Q1 2024

B200 Pricing and Availability

We're tracking 28 cloud providers for the B200, and the pricing spread is significant. While the average sits at $8.53/hr, the lowest price is currently $3.75/hr per GPU (on-demand).

28 providers

Default ranking

Two things group the list before anything else is weighed:

  1. Availability: in stock, then waitlist, then unconfirmed, then out of stock.
  2. Billing type: on-demand rates rank ahead of spot and reserved ones. A spot rate is cheaper because it can be reclaimed, and a reserved rate because it is a commitment.

Within each group, our algorithm weighs four factors to find the most relevant matches for you:

  1. Location: We blend datacenter proximity with provider HQ location.
  2. Price: We blend total hourly price with price-per-GPU to balance affordability and value.
  3. Specs: A light tiebreaker favoring higher GPU count, GPU memory, CPU, and RAM.
  4. Provider diversity: Each time a provider appears in the list, their subsequent offerings are ranked a little lower, so one provider's offerings don't crowd out the top positions.

The two views

By provider groups this ranking one card per provider, and each provider takes the position its best-ranked configuration reached. By configuration lists the same offerings flat, one row each.

In the configuration view you can click any column header to sort by that column, switch billing type with the tabs above the table, and narrow by GPU count, vCPUs or RAM with the filters beneath them. Custom sorting replaces the default relevance ranking.

Transparency and funding

Ads and sponsors: Any paid placements are fixed at the top and labeled as sponsored content.

Affiliates: Any affiliate links will be indicated to you as well. We may earn a commission if you click them, but this never influences the ranking order.

Vast.ai logo

Vast.ai In stock Our sponsor

USA 9 configs 1x-8x

On-Demand from $7.75 Reserved from $9.47 Spot from $5.00
From $7.75 / GPU / hr On-Demand View pricing
Packet·ai logo

Packet·ai In stock

USA 5 configs 1x-8x

On-Demand from $3.75 Custom on request
From $3.75 / GPU / hr On-Demand View pricing
Runpod logo

Runpod In stock

USA 5 configs 1x-8x

On-Demand from $6.79 Reserved on request
From $6.79 / GPU / hr On-Demand SXM View pricing
CoreWeave logo

CoreWeave In stock

USA 2 configs 8x

On-Demand from $8.60 Spot from $4.26
From $8.60 / GPU / hr On-Demand View pricing
GPU.ai logo

GPU.ai In stock

UAE 4 configs 1x-8x

On-Demand from $6.79
From $6.79 / GPU / hr On-Demand SXM View pricing
Spheron logo

Spheron In stock

Singapore 3 configs 1x-8x

On-Demand from $9.36 Spot from $5.34
From $9.36 / GPU / hr On-Demand SXM View pricing
Amazon Web Services logo

AWS In stock

USA 1 config 8x

On-Demand from $14.24
From $14.24 / GPU / hr On-Demand View pricing
Google Cloud logo

Google Cloud In stock

USA 4 configs 8x

On-Demand from $16.11 Reserved from $7.09 Spot from $4.08
From $16.11 / GPU / hr On-Demand View pricing
Lyceum logo

Lyceum In stock

Germany 9 configs 1x-8x

On-Demand from $6.49 Reserved on request Spot from $2.40
Pricing On request Reservation View pricing
Koyeb logo

Koyeb Waitlist

France 1 config 1x

On-Demand from $5.50
From $5.50 / GPU / hr On-Demand View pricing
Fal.ai logo

Fal.ai

USA 2 configs 1x

On-Demand from $6.25 Custom from $3.49
From $6.25 / GPU / hr On-Demand View pricing
Together logo

Together

USA 4 configs 1x

On-Demand from $8.19 Reserved from $6.79
From $8.19 / GPU / hr On-Demand SXM View pricing
Hyperstack logo

Hyperstack

UK 2 configs 1x

On-Demand from $6.00 Reserved from $5.10
From $6.00 / GPU / hr On-Demand View pricing
Nebius logo

Nebius

Netherlands 2 configs 1x

On-Demand from $7.15 Spot from $3.95
From $7.15 / GPU / hr On-Demand SXM View pricing
Beyond.pl logo

Beyond.pl

Poland 7 configs 1x-8x

On-Demand from $5.10 Reserved on request Custom on request
From $5.10 / GPU / hr On-Demand View pricing
Oracle Cloud logo

Oracle Cloud

USA 1 config 8x

On-Demand from $14.00
From $14.00 / GPU / hr On-Demand View pricing
Cirrascale logo

Cirrascale

USA 4 configs 8x

Reserved from $4.86
From $4.86 / GPU / hr Reserved (12mo) View pricing
Civo logo

Civo

UK 3 configs 8x

Reserved from $3.79
From $3.79 / GPU / hr Reserved (36mo) View pricing
Impossible Cloud logo

Impossible Cloud

Germany 1 config 8x

Reserved from $5.58
From $5.58 / GPU / hr Reserved (6mo) View pricing
Green AI Cloud logo

Green AI Cloud

Sweden 1 config 8x

Reserved from $3.35
From $3.35 / GPU / hr Reservation View pricing
GPUaaS logo

GPUaaS

USA 1 config 8x

On-Demand on request
Pricing On request On-Demand View pricing
Canopy Wave logo

Canopy Wave

USA 3 configs 2x-8x

Custom on request
Pricing On request Custom SXM View pricing
CUDO Compute logo

CUDO

UK 1 config 1x

Custom on request
Pricing On request Custom SXM View pricing
Lambda Labs logo

Lambda Labs Out of stock

USA 3 configs 1x-8x

On-Demand from $6.69
From $6.69 / GPU / hr On-Demand SXM View pricing
UpCloud logo

UpCloud Out of stock

Finland 8 configs 1x-8x

On-Demand from $5.20 Spot from $3.90
From $5.20 / GPU / hr On-Demand View pricing
Massed Compute logo

Massed Compute Out of stock

USA 1 config 8x

On-Demand from $5.43
From $5.43 / GPU / hr On-Demand SXM View pricing
Verda logo

Verda Out of stock

Finland 28 configs 1x-8x

On-Demand from $6.11 Reserved from $4.58 Spot from $3.06
From $6.11 / GPU / hr On-Demand SXM View pricing
Vultr logo

Vultr Out of stock

USA 1 config 8x

Spot from $3.20
From $3.20 / GPU / hr Spot View pricing

Heads up: We do our best to keep these specs & prices accurate. However, cloud costs may fluctuate based on region, usage, and other factors not listed here. These are estimates based on common setups and are for informational purposes only. Always verify current rates & exact specs with the provider before provisioning.

Frequently Asked Questions

Why choose the B200?

192GB HBM3e with 1800 GB/s NVLink. Blackwell FP4 Tensor Cores for significantly improved inference throughput over H100. Designed for training and serving the largest foundation models.

When is the B200 not a good fit?

Premium pricing and limited availability as the latest generation. If your workload fits on an H100 or H200, the cost difference may not be justified.

Are B200 prices going up or down?

The median on-demand price across providers has risen about 24% since August 2025, from $5.75 to $7.15/hr per GPU. This reflects the market-wide median, which moves both when providers change prices and when lower or higher priced offerings enter the market.

What size AI models can the B200 run?

With 192GB of VRAM, the B200 can usually run 70B-class models with headroom, and may handle much larger models in 4-bit quantized form depending on runtime overhead, KV cache, and context length.

How much VRAM does the B200 have?

The B200 has 192GB of VRAM. Multi-GPU setups increase total memory, but that memory is not automatically pooled across GPUs.

What is the B200's memory bandwidth?

The B200 has 8,000 GB/s of memory bandwidth. Higher bandwidth helps with faster data transfer between GPU memory and compute cores.

What data types does the B200 support?

The B200 supports 9 precision formats. Training: BF16, FP16, TF32, FP32. Inference: FP4, FP6, FP8, INT8. Scientific: FP64.

Does the B200 support NVLink?

Yes. The B200 supports NVLink with 1800 GB/s of bidirectional bandwidth. This helps accelerate multi-GPU communication.

How much does the B200 cost per hour?

B200 pricing currently ranges from $3.35/hr to $16.11/hr per GPU, depending on the provider, instance type, and billing model.

How much does the B200 cost per month?

At 720 hours per month, one B200 can cost between $2,412.00 to $11,599.20 per month, depending on the provider. Reserved and spot pricing can lower that further.

Which cloud providers offer the B200?

The B200 is available from 28 cloud providers, including Verda, Lyceum, UpCloud. Pricing and availability vary by region and billing model.

Can I rent the B200 in the cloud?

Yes. We currently track 116 B200 listings across 28 cloud providers:

Billing type Listings Avg $/GPU/hr
On-demand 42 $7.23/hr
Reserved 43 $5.85/hr
Spot 22 $4.01/hr
Custom contract 9 $3.49/hr

Technical Specifications

Form Factor 8x NVIDIA B200 SXM
FP4 Tensor Core¹ 144 PFLOPS
FP8/FP6 Tensor Core¹ 72 PFLOPS
INT8 Tensor Core¹ 72 POPS
FP16/BF16 Tensor Core¹ 36 PFLOPS
TF32 Tensor Core¹ 18 PFLOPS
FP32 640 TFLOPS
FP64 320 TFLOPS
FP64 Tensor Core 320 TFLOPS
Memory Up to 1.5TB
NVLink Fifth generation
NVIDIA NVSwitch™ Fourth generation
NVSwitch GPU-to-GPU Bandwidth 1.8TB/s
Total Aggregate Bandwidth 14.4TB/s

¹ With sparsity.

Source: official Nvidia B200 datasheet.

Alternatives to Nvidia B200

Last updated