NVIDIAAPI Host BenchmarksEstimated
Nemotron 3 Ultra 550B — 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
1775 t/s
Lowest latency (TTFT)
90 ms P50
Best price
$0.27 / $1.38
Blended Pricing Workload Calculator
Switch between standard conversational ratios, agentic caching loops, and heavy RAG workloads.
| Provider & Host | Hardware Stack | Output Speed (TPS) | Latency TTFT (P50) | Input Price ($/1M) | Output Price ($/1M) | Cache Read ($/1M) | Blended Price ($/1M) | SLA |
|---|---|---|---|---|---|---|---|---|
NVIDIAOfficial | Cloud TPU / Nvidia H100 | 142 t/s | 2.15 s (3.01 s p90) | $0.6 | $2.75 | $0.06 | $1.14 | 99.99% |
| Groq LPU v2 (SRAM) | 596 t/s | 90 ms (120 ms p90) | $0.51 | $2.34 | $0.048 | $0.97 | 99.95% | |
| CS-3 Wafer-Scale Engine | 1775 t/s | 120 ms (160 ms p90) | $0.54 | $2.48 | — | $1.02 | 99.90% | |
| Nvidia H100 SXM5 | 213 t/s | 1.72 s (2.37 s p90) | $0.39 | $1.92 | $0.036 | $0.77 | 99.95% | |
| Nvidia H100 / L40S | 185 t/s | 1.94 s (2.69 s p90) | $0.27 | $1.38 | $0.03 | $0.55 | 99.90% | |
| Nvidia B200 / H100 | 270 t/s | 1.50 s (2.04 s p90) | $0.36 | $1.79 | $0.036 | $0.72 | 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.