Two Nvidia data center GPUs, one Hopper and one Pascal: 96 GB HBM3 against 8 GB GDDR5 and 21x the memory bandwidth on the H20.
At a glance
Memory
- H20
- 96 GB
- P4
- 8 GB
12x more on H20
Memory bandwidth
- H20
- 4,000 GB/s
- P4
- 192 GB/s
21x higher on H20
Cheapest on-demand
- H20
- --
- P4
- $0.65 /GPU/hr
Price comparison
Median price per GPU per hour by billing type, with the cheapest listing under it.
| Billing | H20 | P4 | Difference |
|---|---|---|---|
| On-demand | No on-demand listings | -- /GPU/hr from $0.65 | |
| Reserved | No reserved listings | -- /GPU/hr from $0.34 (36 mo) | |
| Spot | No spot listings | -- /GPU/hr from $0.12 |
Heads up: A provider's own page may quote a different figure, for example on tax, region or a promotion. A monthly-only plan shows a derived hourly rate. Medians are across providers, one vote each, so a large catalog does not outweigh a small one. No median for on-demand, reserved and spot on the P4: we only show one when at least 3 providers list the GPU on that billing type. Verify before provisioning. More on how we price.
Price history
H20 vs P4 price history
Weekly median on-demand price per GPU per hour
Specs comparison
| Spec | H20 | P4 | Difference |
|---|---|---|---|
| Memory | |||
| Capacity | 96 GB HBM3 | 8 GB GDDR5 | 12x more on H20 |
| Bandwidth | 4,000 GB/s | 192 GB/s | 21x higher on H20 |
| Bus width | 256-bit | ||
| Platform | |||
| Interconnect | PCIe only | PCIe only | |
| Board power | 400 W | 75 W | 5.3x higher on H20 |
| Released | Q1 2024 | Q3 2016 | 7 years newer on H20 |
| Precision | |||
| Supported formats | FP8FP16BF16TF32FP32FP64INT8 | FP16FP32INT8 | |
Source: P4 datasheet.
Where to rent them
Providers listing these GPUs, those with both first, so you can switch without moving clouds.
| Provider | H20 /GPU/hr | P4 /GPU/hr |
|---|---|---|
Google Cloud
|
-- | On-Demand $0.65 |
Similar GPUs
Other comparisons in the same class.
Our data for Nvidia H20 was last updated on Sept. 7, 2026, and for Nvidia P4 on Sept. 7, 2026.