RTX 4500 ADA 24GB vs RTX 5090
The cheapest RTX 5090 ($0.22) costs 32% less per hour than the cheapest RTX 4500 ADA 24GB ($0.32).
Per unit of performance, the RTX 5090 is 68% cheaper: $0.24 versus $0.73 per A100-equivalent hour.
The RTX 5090 has 32 GB of memory against 24 GB on the RTX 4500 ADA 24GB.
| Measure | NVIDIA RTX 4500 Ada Generation 24GB | NVIDIA GeForce RTX 5090 |
|---|---|---|
| Lowest per GPU-hour | $0.32 | $0.22 |
| Median per GPU-hour | $0.81 | $0.87 |
| On-demand from | $0.46 | $0.34 |
| Spot from | — | $0.25 |
| Reserved from | $0.32 | $0.22 |
| Lowest per A100-equivalent hour | $0.73 | $0.24 |
| Live offers | 7 | 405 |
| Providers | 5 | 84 |
| Memory | 24 GB GDDR6 | 32 GB GDDR7 |
| Memory bandwidth | 432 GB/s | 1,792 GB/s |
| FP16 / BF16 | 158.5 TFLOPS | 209.5 TFLOPS |
| FP8 | 317 TFLOPS | 419 TFLOPS |
| Performance vs A100 80GB | 0.44× | 0.92× |
| Architecture | Ada Lovelace | Blackwell |
| Power | 210 W | 575 W |