Every host for DeepSeek V4 Flash 0731
| Host | Blended | Input | Output | Cached | Precision | Uptime |
|---|---|---|---|---|---|---|
| Relace | $0.075 | $0.048 | $0.157 | $0.028 | fp4 | 100.0% |
| DeepInfra | $0.09 | $0.06 | $0.18 | $0.015 | fp8 | 100.0% |
| Sail | $0.113 | $0.09 | $0.18 | $0.02 | — | — |
| Digitalocean | $0.149 | $0.119 | $0.238 | $0.024 | — | 100.0% |
| Sail Research | $0.15 | $0.15 | $0.15 | $0.042 | fp4 | 100.0% |
| Baseten | $0.163 | $0.13 | $0.26 | $0.028 | fp32 | 99.6% |
| Perplexity | $0.163 | $0.13 | $0.26 | $0.028 | — | — |
| Venice | $0.163 | $0.13 | $0.26 | $0.028 | — | 99.7% |
| Wafer | $0.163 | $0.14 | $0.23 | $0.135 | nvfp4 | 100.0% |
| CoreWeave | $0.168 | $0.13 | $0.28 | $0.07 | fp8 | 100.0% |
| Weights & Biases | $0.168 | $0.13 | $0.28 | $0.07 | — | — |
| Reka | $0.17 | $0.05 | $0.528 | $0.02 | — | 100.0% |
| Cohere | $0.175 | $0.14 | $0.28 | $0.07 | — | 100.0% |
| Nebius | $0.175 | $0.14 | $0.28 | — | — | — |
| Parasail | $0.175 | $0.14 | $0.28 | $0.05 | fp8 | 99.7% |
| Together | $0.175 | $0.14 | $0.28 | $0.03 | — | 100.0% |
| Inceptron | $0.186 | $0.11 | $0.413 | $0.027 | fp4 | 99.8% |
| Streamlake | $0.198 | $0.132 | $0.396 | $0.0042 | fp8 | 99.9% |
| Inferenceapis | $0.238 | $0.19 | $0.38 | $0.04 | — | — |
| Alibaba Cloud | $0.25 | $0.20 | $0.40 | $0.04 | — | — |
| Mancer 2 | $0.30 | $0.20 | $0.60 | — | fp8 | 99.8% |
| Fireworks | $0.33 | $0.22 | $0.66 | $0.007 | — | — |
| SiliconFlow | $0.33 | $0.22 | $0.66 | $0.028 | fp8 | 100.0% |
| GMI Cloud | $0.429 | $0.286 | $0.858 | $0.0091 | fp8 | 100.0% |
| Phala | $0.462 | $0.308 | $0.924 | $0.02 | — | 100.0% |
| Scaleway | $0.50 | $0.40 | $0.80 | $0.08 | — | — |
| Alibaba Cloud | $0.528 | $0.352 | $1.06 | $0.035 | — | 99.6% |
| Novita | $0.614 | $0.409 | $1.23 | $0.026 | fp8 | 100.0% |
| Atlas Cloud | $0.66 | $0.44 | $1.32 | $0.028 | fp4 | 100.0% |
| Azure | $0.66 | $0.44 | $1.32 | $0.014 | — | — |
| Baidu | $0.66 | $0.44 | $1.32 | $0.014 | fp8 | 100.0% |
| Cloudflare | $0.66 | $0.44 | $1.32 | $0.014 | — | 99.9% |
Cheapest price over time
- Input
- Output
What your workload costs
- Relace$29.93
- DeepInfra$36.00
- Sail$45.00
- Digitalocean$59.50
- Sail Research$60.00
- Baseten$65.00
- Perplexity$65.00
- Venice$65.00
- Wafer$65.00
- CoreWeave$67.00
- Weights & Biases$67.00
- Reka$67.80
20 more hosts cost more.
Rent GPUs or pay per token?
- Renting the GPUs
- $16,878/mo
- Break-even volume
- 226B tokens/mo
- Needed around the clock
- 85.8K tok/s
8 × H100 cost $16,878 a month. The cheapest API (Relace) charges $0.075 per 1M blended tokens, so the GPUs only pay off above 226B tokens a month: the box must serve 85.8K tokens a second, every second.
GPU price: cheapest on-demand offer per GPU-hour × 730 hours. No throughput is assumed: it depends on the model size, precision, batch size and serving software, so measure yours. Power, storage, egress and engineering time are not included.
More from DeepSeek
Quality data: LMArena (CC BY 4.0), Epoch AI (CC BY 4.0). Prices: OpenRouter, LiteLLM (MIT), provider pages. List prices in US dollars before discounts and taxes; other currencies converted at the European Central Bank reference rate. The figures on these pages are free to reuse under CC BY 4.0 with a link to this page.