RTX 4000 ADA 20GB vs RTX 5090
The cheapest RTX 4000 ADA 20GB ($0.067) costs 69% less per hour than the cheapest RTX 5090 ($0.22).
Per unit of performance, the RTX 5090 is 2% cheaper: $0.24 versus $0.24 per A100-equivalent hour.
The RTX 5090 has 32 GB of memory against 20 GB on the RTX 4000 ADA 20GB.
| Measure | NVIDIA RTX 4000 Ada Generation 20GB | NVIDIA GeForce RTX 5090 |
|---|---|---|
| Lowest per GPU-hour | $0.067 | $0.22 |
| Median per GPU-hour | $0.47 | $0.87 |
| On-demand from | $0.067 | $0.34 |
| Spot from | — | $0.25 |
| Reserved from | $0.18 | $0.22 |
| Lowest per A100-equivalent hour | $0.24 | $0.24 |
| Live offers | 61 | 405 |
| Providers | 24 | 84 |
| Memory | 20 GB GDDR6 | 32 GB GDDR7 |
| Memory bandwidth | 360 GB/s | 1,792 GB/s |
| FP16 / BF16 | 81.9 TFLOPS | 209.5 TFLOPS |
| FP8 | 163.8 TFLOPS | 419 TFLOPS |
| Performance vs A100 80GB | 0.28× | 0.92× |
| Architecture | Ada Lovelace | Blackwell |
| Power | 130 W | 575 W |