Two Nvidia workstation GPUs, one Maxwell and one Ada Lovelace: 2 GB DDR3 against 24 GB GDDR6 and 15x the memory bandwidth on the RTX 4500 Ada.
At a glance
Memory
- K620
- 2 GB
- RTX 4500 Ada
- 24 GB
12x more on RTX 4500 Ada
Memory bandwidth
- K620
- 29 GB/s
- RTX 4500 Ada
- 432 GB/s
15x higher on RTX 4500 Ada
Cheapest on-demand
- K620
- --
- RTX 4500 Ada
- --
Price comparison
Median price per GPU per hour by billing type, with the cheapest listing under it.
Price history
K620 vs RTX 4500 Ada price history
Weekly median on-demand price per GPU per hour
Specs comparison
| Spec | K620 | RTX 4500 Ada | Difference |
|---|---|---|---|
| Memory | |||
| Capacity | 2 GB DDR3 | 24 GB GDDR6 | 12x more on RTX 4500 Ada |
| Bandwidth | 29 GB/s | 432 GB/s | 15x higher on RTX 4500 Ada |
| Bus width | 128-bit | 192-bit | 1.5x wider on RTX 4500 Ada |
| Compute · dense | |||
| FP8 | Not supported | 317 TFLOPS | RTX 4500 Ada only |
| FP16 / BF16 | Not supported | 158.5 TFLOPS | RTX 4500 Ada only |
| FP32 | 0.9 TFLOPS | 39.6 TFLOPS | 46x higher on RTX 4500 Ada |
| Platform | |||
| Interconnect | PCIe only | PCIe only | |
| Board power | 45 W | 210 W | 4.7x higher on RTX 4500 Ada |
| Released | Q3 2014 | Q3 2023 | 9 years newer on RTX 4500 Ada |
| Precision | |||
| Supported formats | FP32 | FP8FP16BF16TF32FP32INT4INT8 | |
Compute: Vendor peak figures, dense and per GPU. Not measured throughput.
Source: K620 datasheet and RTX 4500 Ada datasheet.
Similar GPUs
Other comparisons in the same class.
Our data for Nvidia K620 was last updated on Sept. 7, 2026, and for Nvidia RTX 4500 Ada on Sept. 7, 2026.