Nvidia A5000 vs Nvidia H100

A workstation GPU against a data center card, one Ampere and one Hopper: 24 GB GDDR6 against 80 GB HBM3 and 4.4x the memory bandwidth on the H100.

At a glance

Updated 6 minutes ago

Nvidia A5000

Nvidia H100

Cheapest listed price per GPU per hour by billing type, and the median across providers.

Billing	A5000	H100	Difference
On-demand	$0.25	$1.50	6x more expensive on H100
Spot	$0.09	$0.13	33% cheaper on A5000
Reserved	$0.25	$1.53	6.1x more expensive on H100
Median	$1.38	$3.39	2.5x more expensive on H100

Median: Median across providers, one vote each, so a large catalog does not outweigh a small one.

Weekly median on-demand price per GPU per hour

Last 12 months

Aggregating historical prices...

Spec	A5000	H100	Difference
Memory
Capacity	24 GB GDDR6	80 GB HBM3 94 GB on NVL	3.3x more on H100
Bandwidth	768 GB/s	3,350 GB/s 2,000 GB/s on PCIe 3,938 GB/s on NVL	4.4x higher on H100
Bus width	384-bit	5,120-bit	13x wider on H100
Compute · dense
FP8	Not supported	1,979 TFLOPS	H100 only
FP16 / BF16	111.1 TFLOPS	989.5 TFLOPS	8.9x higher on H100
INT8		1,979 TOPS
FP32	27.8 TFLOPS	67 TFLOPS	2.4x higher on H100
FP64	Not supported	34 TFLOPS	H100 only
Platform
Interconnect	NVLink 112 GB/s	NVLink 900 GB/s NVLink 600 GB/s on PCIe, NVL	8x higher on H100
Board power	230 W	700 W	3x higher on H100
Released	Q2 2021	Q3 2022	17 months newer on H100
Precision
Supported formats	FP16BF16TF32FP32INT4INT8	FP8FP16BF16TF32FP32FP64INT8

Figures are for each model's headline configuration. Where a form factor carries a different number, it is named under the figure.

Compute: Vendor peak figures, dense and per GPU. Not measured throughput.

Providers listing these GPUs, those with both first, so you can switch without moving clouds.