Two Nvidia data center GPUs, one Ampere and one Pascal: 16 GB GDDR6 against 8 GB GDDR5 and memory bandwidth within 5%.
At a glance
Updated 4 minutes ago- Memory
- 16 GB
- GDDR6
- Memory bandwidth
- 200 GB/s
- From
- $0.47 /GPU/hr
- 3 providers
- Memory
- 8 GB
- GDDR5
- Memory bandwidth
- 192 GB/s
- From
- $0.65 /GPU/hr
- 1 provider
Price comparison
Cheapest listed price per GPU per hour by billing type, and the median across providers.
| Billing | A16 | P4 | Difference |
|---|---|---|---|
| On-demand | $0.47 | $0.65 | 28% cheaper on A16 |
| Spot | No spot listings | $0.12 | |
| Reserved | No reserved listings | $0.34 | |
| Median | $0.56 | $0.81 | 31% cheaper on A16 |
Median: Median across providers, one vote each, so a large catalog does not outweigh a small one.
Price history
A16 vs P4 price history
Weekly median on-demand price per GPU per hour
Specs comparison
| Spec | A16 | P4 | Difference |
|---|---|---|---|
| Memory | |||
| Capacity | 16 GB GDDR6 | 8 GB GDDR5 | 2x more on A16 |
| Bandwidth | 200 GB/s | 192 GB/s | Within 5% |
| Bus width | 128-bit | 256-bit | 2x wider on P4 |
| Compute · dense | |||
| FP16 / BF16 | 17.9 TFLOPS | ||
| INT8 | 35.9 TOPS | 22 TOPS | 1.6x higher on A16 |
| FP32 | 4.5 TFLOPS | 5.5 TFLOPS | 1.2x higher on P4 |
| Platform | |||
| Interconnect | PCIe only | PCIe only | |
| Released | Q2 2021 | Q3 2016 | 5 years newer on A16 |
| Precision | |||
| Supported formats | FP16BF16TF32FP32INT4INT8 | FP16FP32INT8 | |
Compute: Vendor peak figures, dense and per GPU. Not measured throughput.
Source: A16 datasheet and P4 datasheet.
Where to rent them
Providers listing these GPUs, those with both first, so you can switch without moving clouds.
| Provider | A16 /GPU/hr | P4 /GPU/hr |
|---|---|---|
Vultr
|
On-Demand $0.47 | -- |
Runcrate
|
On-Demand $0.56 | -- |
Sesterce
|
On-Demand $0.56 | -- |
Google Cloud
|
-- | On-Demand $0.65 |
Similar GPUs
Other comparisons in the same class.
Our data for Nvidia A16 was last updated on Sept. 7, 2026, and for Nvidia P4 on Sept. 7, 2026.