B200 vs V100 32GB
The cheapest V100 32GB ($0.017) costs 98% less per hour than the cheapest B200 ($0.99).
Per unit of performance, the V100 32GB is 72% cheaper: $0.04 versus $0.14 per A100-equivalent hour.
The B200 has 180 GB of memory against 32 GB on the V100 32GB.
| Measure | NVIDIA B200 180GB SXM | NVIDIA Tesla V100 32GB |
|---|---|---|
| Lowest per GPU-hour | $0.99 | $0.017 |
| Median per GPU-hour | $7.89 | $2.32 |
| On-demand from | $0.99 | $0.017 |
| Spot from | $1.16 | $0.51 |
| Reserved from | $3.89 | $0.19 |
| Lowest per A100-equivalent hour | $0.14 | $0.04 |
| Live offers | 146 | 193 |
| Providers | 63 | 43 |
| Memory | 180 GB HBM3e | 32 GB HBM2 |
| Memory bandwidth | 8,000 GB/s | 900 GB/s |
| FP16 / BF16 | 2,250 TFLOPS | 125 TFLOPS |
| FP8 | 4,500 TFLOPS | — |
| Performance vs A100 80GB | 6.96× | 0.42× |
| Architecture | Blackwell | Volta |
| Power | 1,000 W | 300 W |