The H800 is newer, with 40x the memory and 69x the bandwidth.
At a glance
Memory
40x more on H800
- H800
- 80 GB
- K620
- 2 GB
Memory bandwidth
69x higher on H800
- H800
- 2,000 GB/s
- K620
- 29 GB/s
Cheapest on-demand
- H800
- --
- K620
- --
Price comparison
Median price per GPU per hour by billing type, with the cheapest listing under it.
Price history
H800 vs K620 price history
Weekly median on-demand price per GPU per hour
Aggregating historical prices...
Specs comparison
| Spec | H800 | K620 | Difference |
|---|---|---|---|
| Memory | |||
| Capacity | 80 GB HBM2e | 2 GB DDR3 | 40x more on H800 |
| Bandwidth | 2,000 GB/s | 29 GB/s | 69x higher on H800 |
| Bus width | 5,120-bit | 128-bit | 40x wider on H800 |
| Compute · dense | |||
| FP8 | 1,513 TFLOPS | Not supported | H800 only |
| FP16 / BF16 | 756.5 TFLOPS | Not supported | H800 only |
| INT8 | 1,513 TOPS | Not supported | H800 only |
| FP32 | 51 TFLOPS | 0.9 TFLOPS | 59x higher on H800 |
| Platform | |||
| Interconnect | NVLink 400 GB/s | PCIe only | NVLink on H800 only |
| Board power | 350 W | 45 W | 7.8x higher on H800 |
| Released | Q1 2023 | Q3 2014 | 9 years newer on H800 |
| Precision | |||
| Supported formats | FP8FP16BF16TF32FP32FP64INT8 | FP32 | |
Compute: Vendor peak figures, dense and per GPU. Not measured throughput.
Source: K620 datasheet.
Similar GPUs
Other comparisons in the same class.
-
Nvidia H100
80 GB HBM3 · 68% more bandwidth than the H800
From $1.73 /GPU/hr Compare -
Nvidia H200
141 GB HBM3e · 2.4x the bandwidth of the H800
From $2.09 /GPU/hr Compare
-
Nvidia A800
40 GB HBM2 · 22% less bandwidth than the H800
From $1.18 /GPU/hr Compare
Our data for Nvidia H800 was last updated on Sept. 14, 2026, and for Nvidia K620 on Sept. 14, 2026.