What's the price difference between NVIDIA B200 and NVIDIA H200?

On Runcrate today, NVIDIA B200 starts at $7.93/hr and NVIDIA H200 starts at $4.63/hr; NVIDIA H200 is cheaper by 42%.

Can I deploy NVIDIA B200 or NVIDIA H200 on Runcrate today?

NVIDIA B200 is in stock as a reserved cluster: we verify your company and your need, then take your exact spec to GPU owners that fit, and you see their offers side by side. NVIDIA H200 is on demand in 5 regions from $4.63/hr, billed by the second on most GPUs.

How much VRAM does each GPU have?

NVIDIA B200 has 180GB HBM3e. NVIDIA H200 has 141GB HBM3e.

B200 or H200: which holds a 70B model in 8-bit weights?

It needs 70 GB of 8-bit weights. On the B200: 1 GPU, 110 GB spare. On the H200: 1 GPU, 71 GB spare. Pick the B200: same GPU count, 67% more memory bandwidth.

GPU comparison

B200 vs H200.

The B200 is ahead on memory (180 GB vs 141 GB) and speed (2.3x the FP16 throughput and 67% more memory bandwidth). The H200 makes sense when your model fits in 141 GB and its hourly price is lower.

Launch H200 Reserve a cluster

Rent either on Runcrate

B200from $7.93/hr
$0.044 per GB/hr$0.35 per 100 TFLOPS/hr
Reserve a cluster
H200Cheaperfrom $4.63/hr
$0.033 per GB/hr$0.47 per 100 TFLOPS/hr
Launch H200

Live hourly price, then per GB of memory and per 100 FP16 TFLOPS.

Which is better for

Which to pick for each job. Memory fit and throughput, from the datasheets.

JobB200H200Pick

Hold a 70B model in 8-bit weights
70 GB of 8-bit weights
B200: 1 GPU, 110 GB spare
H200: 1 GPU, 71 GB spare
B200. Same GPU count, 67% more memory bandwidth.
Serve an 8B model in 16-bit weights
16 GB of 16-bit weights
B200: 1 GPU, 164 GB spare
H200: 1 GPU, 125 GB spare
B200. Same GPU count, 67% more memory bandwidth.
Fine-tune a 70B model with QLoRA
About 46 GB with 4-bit weights
B200: 1 GPU, 134 GB spare
H200: 1 GPU, 95 GB spare
B200. Same GPU count, 2.3x the FP16 throughput.
Train on a full node
FP16 tensor throughput per node
B200: 8x, 18,000 TFLOPS, NVLink
H200: 8x, 7,912 TFLOPS, NVLink
B200. 2.3x the FP16 throughput per GPU.
Throughput per watt
FP16 TFLOPS per 100 W of TDP
B200: 225 TFLOPS per 100 W
H200: 141 TFLOPS per 100 W
B200. 59% more FP16 per watt.

Compare another pair

Blackwell

NVIDIA B200

from$7.93/hr

Reserved clusters

Memory: 180 GB HBM3e
FP16 tensor: 2250 TFLOPS
Bandwidth: 8.0 TB/s

Reserve a cluster Deploy B200 now

Full pricing details

Hopper

NVIDIA H200

from$4.63/hrCheaper

Available now · 5 regions

Memory: 141 GB HBM3e
FP16 tensor: 989 TFLOPS
Bandwidth: 4.8 TB/s

Launch H200 Reserve a cluster

Full pricing details

At a glance

Spec by spec. The longer bar is the better number.

Memory

NVIDIA B200180 GB

NVIDIA H200141 GB

Memory bandwidth

NVIDIA B2008.0 TB/s

NVIDIA H2004.8 TB/s

FP16 tensor

NVIDIA B2002250 TFLOPS

NVIDIA H200989 TFLOPS

Hourly price, from (live)

NVIDIA B200$7.93/hr

NVIDIA H200$4.63/hr

Full specs

The complete spec sheet.

Spec	NVIDIA B200	NVIDIA H200
Architecture	Blackwell	Hopper
Memory	180 GB	141 GB
Memory type	HBM3e	HBM3e
Memory bandwidth	8.0 TB/s	4.8 TB/s
FP32	75 TFLOPS	67 TFLOPS
FP16 tensor	2250 TFLOPS	989 TFLOPS
INT8	4500 TOPS	2680 TOPS
TDP	1000 W	700 W
Form factor	SXM	SXM
NVLink	Yes	Yes
GPUs per node	8	8
Hourly price (live)	from $7.93/hr	from $4.63/hr

Questions

B200 or H200. What buyers ask first.

Other comparisons