Two Nvidia data center GPUs, one Ampere and one Blackwell: 80 GB HBM2e against 192 GB HBM3e and 4.1x the memory bandwidth on the B100.
At a glance
Memory
- A100
- 80 GB
- B100
- 192 GB
2.4x more on B100
Memory bandwidth
- A100
- 1.94 TB/s
- B100
- 8.00 TB/s
4.1x higher on B100
Cheapest on-demand
- A100
- $0.38 /GPU/hr
- B100
- --
Price comparison
Median price per GPU per hour by billing type, with the cheapest listing under it.
| Billing | A100 | B100 | Difference |
|---|---|---|---|
| On-demand | $1.76 /GPU/hr from $0.38 | No on-demand listings | |
| Reserved | $1.43 /GPU/hr from $0.57 (1 mo) | No reserved listings | |
| Spot | $1.08 /GPU/hr from $0.20 | No spot listings |
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
A100 vs B100 price history
Weekly median on-demand price per GPU per hour
Specs comparison
| Spec | A100 | B100 | Difference |
|---|---|---|---|
| Memory | |||
| Capacity | 80 GB HBM2e 40 GB on SXM, PCIe | 192 GB HBM3e | 2.4x more on B100 |
| Bandwidth | 1,935 GB/s 2,039 GB/s on SXM 1,555 GB/s on SXM, PCIe | 8,000 GB/s | 4.1x higher on B100 |
| Bus width | 5,120-bit | ||
| Compute · dense | |||
| FP4 | Not supported | 7,000 TFLOPS | B100 only |
| FP8 | Not supported | 3,500 TFLOPS | B100 only |
| FP16 / BF16 | 312 TFLOPS | 1,750 TFLOPS | 5.6x higher on B100 |
| INT8 | 624 TOPS | 3,500 TOPS | 5.6x higher on B100 |
| FP32 | 19.5 TFLOPS | 60 TFLOPS | 3.1x higher on B100 |
| FP64 | 9.7 TFLOPS | 30 TFLOPS | 3.1x higher on B100 |
| Platform | |||
| Interconnect | NVLink 600 GB/s | NVLink 1,800 GB/s | 3x higher on B100 |
| Released | Q2 2020 | Q4 2024 | 4 years newer on B100 |
| Precision | |||
| Supported formats | FP16BF16TF32FP32FP64INT4INT8 | FP4FP6FP8FP16BF16TF32FP32FP64INT8 | |
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: A100 datasheet and B100 datasheet.
Where to rent them
Providers listing these GPUs, those with both first, so you can switch without moving clouds.
| Provider | A100 /GPU/hr | B100 /GPU/hr |
|---|---|---|
Vast.ai
|
On-Demand $0.38 | -- |
Lium
|
On-Demand $0.68 | -- |
GPU.ai
|
On-Demand $0.71 | -- |
Runpod
|
On-Demand $1.00 | -- |
Thunder Compute
|
On-Demand $1.09 | -- |
Verda
|
On-Demand $1.25 | -- |
Massed Compute
|
On-Demand $1.35 | -- |
EmpirioLabs AI
|
On-Demand $1.40 | -- |
43 providers list the A100 and 0 the B100. Every listing is on the A100 page and the B100 page.
Similar GPUs
Other comparisons in the same class.
-
Nvidia H100
80 GB HBM3 · 73% more bandwidth than the A100
From $1.30 /GPU/hr Compare -
Nvidia RTX 5090
32 GB GDDR7 · PCIe only
From $0.35 /GPU/hr Compare
-
Nvidia RTX 4090
24 GB GDDR6X · 48% less bandwidth than the A100
From $0.23 /GPU/hr Compare
-
Nvidia B200
180 GB HBM3e · released Q1 2024
From $3.75 /GPU/hr Compare
Our data for Nvidia A100 was last updated on Sept. 14, 2026, and for Nvidia B100 on Sept. 14, 2026.