B200 GPU Rental

Rent B200 GPUs on Lium from $3.70 per GPU-hour, built for large-scale AI training, inference, and HPC. Billed per second, no minimum. Deploy in seconds from the browser or the CLI.
Refer a friend: earn 5% of their top-ups for 3 months, then 2.5% for 3 more. Get your link
Prices and availability as of 02:15 UTC
Price per GPU-hour
$3.70 – $8.89
Live asks across available pods
Smallest rental: 1 GPU at $6.10/hour
Available now
5 pods
29 GPUs in total
Reference price
$5.60
Marketplace baseline per GPU-hour

About the B200

The B200 is NVIDIA’s Blackwell data center GPU and the successor to the H100. Two things matter in practice: memory and low precision. Each card carries 180 GB of HBM3e at 8 TB/s, more than twice the capacity and bandwidth of an H100, and Blackwell adds native FP4 alongside FP8, so inference kernels that can drop precision get another step up in throughput. NVLink runs at 1.8 TB/s between cards in a node. The card draws up to 1,000 W, which is why it turns up in purpose-built eight-GPU systems rather than in retrofitted racks.

Common uses

  • Serving 70B+ models on a single card
  • FP4 and FP8 inference at scale
  • Multi-node training over NVLink
  • Long-context and high-batch workloads

B200 specifications

VRAM180 GB HBM3e
Memory bandwidth8,000 GB/s
FP16 tensor2.25 PFLOPS
FP8 tensor4.5 PFLOPS
InterconnectNVLink 1.8 TB/s
Power1,000 W
ArchitectureBlackwell
Released2024
Tensor figures are NVIDIA’s published dense numbers for the card, without sparsity; sparsity doubles them. What a pod actually delivers depends on the model, the kernels and the host it runs on.

What runs on one B200

The largest open models one 180 GB card holds for inference (weights plus about 20% for activations and cache). Rent several GPUs in one pod to go bigger.
PrecisionLargest model that fits
FP1670B parameters
8-bit70B parameters
4-bit70B parameters
Running one B200 around the clock for a month (730 hours) costs about $2,701 at the lowest live ask; you pay only for the seconds a pod runs.

Frequently asked questions

How much does it cost to rent a B200 on Lium?

$3.70 to $8.89 per GPU-hour right now, on 29 GPUs. Providers set the price.

Where is the cheapest place to rent a B200 by the hour?

On Lium, B200 pods start at $3.70 per GPU-hour right now, and billing is per second with no minimum.

Is there a minimum rental period?

No. Lium bills per second with no minimum and no reservation: start a pod when you need it, delete it when done, and pay for the seconds it ran.

How do I rent a B200 pod?

Top up by card or crypto, pick a B200 pod on the marketplace, choose a Docker template, add your SSH key and deploy. From a script: export LIUM_API_KEY, then run lium up with the node id.

Where do the machines run, and what happens to my data?

Providers run their own hardware in their own facilities, and every pod names its city and country, so you pick the region. Your data stays on the rented machine and is released when you delete the pod; attach a volume for storage that outlives a pod.

Compare with

GPUVRAMPrice per GPU-hour
GB200192 GBreference $5.60/hr
B300 SXM6288 GBfrom $4.00/hr
GB300252 GBreference $8.00/hr
For AI agents (CLI install, agent skill, headless signup): llms-full.txt