Two Nvidia data center GPUs, one Ampere and one Hopper: 16 GB GDDR6 against 80 GB HBM3, 17x the memory bandwidth on the H100, and NVLink 900 GB/s on the H100 where the A16 is PCIe only.
At a glance
Memory
- A16
- 16 GB
- H100
- 80 GB
5x more on H100
Memory bandwidth
- A16
- 200 GB/s
- H100
- 3,350 GB/s
17x higher on H100
Cheapest on-demand
- A16
- $0.47 /GPU/hr
- H100
- $1.50 /GPU/hr
3.2x more expensive on H100
Price comparison
Median price per GPU per hour by billing type, with the cheapest listing under it.
| Billing | A16 | H100 | Difference |
|---|---|---|---|
| On-demand | $0.56 /GPU/hr from $0.47 | $3.39 /GPU/hr from $1.50 | 6.1x more expensive on H100 |
| Reserved | No reserved listings | $3.16 /GPU/hr from $1.53 (1 mo) | |
| Spot | No spot listings | $2.15 /GPU/hr from $0.14 |
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. Verify before provisioning. More on how we price.
Price history
A16 vs H100 price history
Weekly median on-demand price per GPU per hour
Specs comparison
| Spec | A16 | H100 | Difference |
|---|---|---|---|
| Memory | |||
| Capacity | 16 GB GDDR6 | 80 GB HBM3 94 GB on NVL | 5x more on H100 |
| Bandwidth | 200 GB/s | 3,350 GB/s 2,000 GB/s on PCIe 3,938 GB/s on NVL | 17x higher on H100 |
| Bus width | 128-bit | 5,120-bit | 40x wider on H100 |
| Compute · dense | |||
| FP8 | Not supported | 1,979 TFLOPS | H100 only |
| FP16 / BF16 | 17.9 TFLOPS | 989.5 TFLOPS | 55x higher on H100 |
| INT8 | 35.9 TOPS | 1,979 TOPS | 55x higher on H100 |
| FP32 | 4.5 TFLOPS | 67 TFLOPS | 15x higher on H100 |
| FP64 | Not supported | 34 TFLOPS | H100 only |
| Platform | |||
| Interconnect | PCIe only | NVLink 900 GB/s NVLink 600 GB/s on PCIe, NVL | NVLink on H100 only |
| 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.
Source: A16 datasheet and H100 datasheet.
Where to rent them
Providers listing these GPUs, those with both first, so you can switch without moving clouds.
| Provider | A16 /GPU/hr | H100 /GPU/hr |
|---|---|---|
Runcrate
|
On-Demand $0.56 | On-Demand $2.75 |
Vultr
|
On-Demand $0.47 | -- |
Sesterce
|
On-Demand $0.56 | -- |
Lium
|
-- | On-Demand $1.50 |
HyperAI
|
-- | On-Demand $1.80 |
Vast.ai
|
-- | On-Demand $2.00 |
GPU.ai
|
-- | On-Demand $2.00 |
Koyeb
|
-- | On-Demand $2.50 |
Massed Compute
|
-- | On-Demand $2.73 |
Lyceum
|
-- | On-Demand $2.79 |
3 providers list the A16 and 53 the H100. Every listing is on the A16 page and the H100 page.
Similar GPUs
Other comparisons in the same class.
Our data for Nvidia A16 was last updated on Sept. 7, 2026, and for Nvidia H100 on Sept. 7, 2026.