Two Nvidia data center GPUs, one Ampere and one Pascal: 24 GB GDDR6 against 8 GB GDDR5 and 3.1x the memory bandwidth on the A10G.
At a glance
Updated 7 hours ago- Memory
- 24 GB
- GDDR6
- Memory bandwidth
- 600 GB/s
- From
- $1.00 /GPU/hr
- 2 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 | A10G | P4 | Difference |
|---|---|---|---|
| On-demand | $1.00 | $0.65 | 35% cheaper on P4 |
| Spot | No spot listings | $0.12 | |
| Reserved | $0.43 | $0.34 | 21% cheaper on P4 |
| Median | $1.60 | $0.81 | 49% cheaper on P4 |
Median: Median across providers, one vote each, so a large catalog does not outweigh a small one.
Price history
A10G vs P4 price history
Weekly median on-demand price per GPU per hour
Specs comparison
| Spec | A10G | P4 | Difference |
|---|---|---|---|
| Memory | |||
| Capacity | 24 GB GDDR6 | 8 GB GDDR5 | 3x more on A10G |
| Bandwidth | 600 GB/s | 192 GB/s | 3.1x higher on A10G |
| Bus width | 256-bit | ||
| Compute · dense | |||
| FP16 / BF16 | 70 TFLOPS | ||
| INT8 | 140 TOPS | 22 TOPS | 6.4x higher on A10G |
| FP32 | 35 TFLOPS | 5.5 TFLOPS | 6.4x higher on A10G |
| Platform | |||
| Interconnect | PCIe only | PCIe only | |
| Board power | 300 W | 75 W | 4x higher on A10G |
| Released | Q2 2021 | Q3 2016 | 5 years newer on A10G |
| Precision | |||
| Supported formats | FP16BF16TF32FP32INT4INT8 | FP16FP32INT8 | |
Compute: Vendor peak figures, dense and per GPU. Not measured throughput.
Source: A10G datasheet and P4 datasheet.
Where to rent them
Providers listing these GPUs, those with both first, so you can switch without moving clouds.
| Provider | A10G /GPU/hr | P4 /GPU/hr |
|---|---|---|
Hugging Face
|
On-Demand $1.00 | -- |
AWS
|
On-Demand $1.01 | -- |
Google Cloud
|
-- | On-Demand $0.65 |
Similar GPUs
Other comparisons in the same class.
Our data for Nvidia A10G was last updated on Sept. 7, 2026, and for Nvidia P4 on Sept. 7, 2026.