Tok/s vs cost of self-hosting deepseek-ai/DeepSeek-V4.1-Flash on H100:8, H200:8.
Run it yourself
deepseek-v4.1-flash-h100@v2 · H100:8
$ iwant launch deepseek-v4.1-flash-h100deepseek-v4.1-flash@v3 · H200:8
$ iwant launch deepseek-v4.1-flashResults
Dashed lines estimate the same throughput at another cloud's price for the same GPUs.
| setup | $/hour | peak tok/s | $/1M at peak | |
|---|---|---|---|---|
| deepseek-v4.1-flash-h100@v2 · 8x H100 GCP europe-west2-b Spot | $11.45 | 6,030 | $0.5275 | |
| deepseek-v4.1-flash@v3 · 8x H200 GCP asia-south1-b Spot | $60.89 | 11,328 | $1.4931 | |
| deepseek-v4.1-flash-h100@v2 · GCP europe-north1 spot (estimated) | $8.70 | 6,030 | $0.4010 | |
| deepseek-v4.1-flash-h100@v2 · AWS us-east-1 on-demand (estimated) | $55.04 | 6,030 | $2.5353 | |
| deepseek-v4.1-flash@v3 · AWS ap-northeast-2 spot (estimated) | $12.18 | 11,328 | $0.2988 | |
| deepseek-v4.1-flash@v3 · AWS us-east-1 on-demand (estimated) | $63.30 | 11,328 | $1.5521 |