DeepSeekAPI Host BenchmarksEstimated
DeepSeek V4 Flash — API Providers & Speed Matrix
Modelled throughput, TTFT latency and prompt-cache pricing estimates across cloud hosting endpoints, derived from official list prices and public benchmarks.
Fastest provider
1413 t/s
Lowest latency (TTFT)
90 ms P50
Best price
$0.06 / $0.14
Blended Pricing Workload Calculator
Switch between standard conversational ratios, agentic caching loops, and heavy RAG workloads.
| Provider & Host | Hardware Stack | Output Speed (TPS) | Latency TTFT (P50) | Input Price ($/1M) | Output Price ($/1M) | Cache Read ($/1M) | Blended Price ($/1M) | SLA |
|---|---|---|---|---|---|---|---|---|
DeepSeekOfficial | Cloud TPU / Nvidia H100 | 113 t/s | 1.34 s (1.88 s p90) | $0.14 | $0.28 | $0.014 | $0.18 | 99.99% |
| Groq LPU v2 (SRAM) | 475 t/s | 90 ms (120 ms p90) | $0.12 | $0.24 | $0.011 | $0.15 | 99.95% | |
| CS-3 Wafer-Scale Engine | 1413 t/s | 120 ms (160 ms p90) | $0.13 | $0.25 | — | $0.16 | 99.90% | |
| Nvidia H100 SXM5 | 170 t/s | 1.07 s (1.47 s p90) | $0.09 | $0.2 | $0.008 | $0.12 | 99.95% | |
| Nvidia H100 / L40S | 147 t/s | 1.21 s (1.68 s p90) | $0.06 | $0.14 | $0.007 | $0.08 | 99.90% | |
| Nvidia B200 / H100 | 215 t/s | 940 ms (1.27 s p90) | $0.08 | $0.18 | $0.008 | $0.1 | 99.99% |
Figures are modelled estimates based on published list prices and public throughput benchmarks — not measured on live endpoints. Verify current pricing and terms with each host before committing.