IBMAPI Host BenchmarksEstimated
Granite 4.2 3B Instruct — 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
1850 t/s
Lowest latency (TTFT)
90 ms P50
Best price
$0.04 / $0.1
Blended Pricing Workload Calculator
Switch between standard conversational ratios, agentic caching loops, and heavy RAG workloads.
| Provider & Host | Hardware Stack | Output Speed (TPS) | Latency TTFT (P50) | Input Price ($/1M) | Output Price ($/1M) | Cache Read ($/1M) | Blended Price ($/1M) | SLA |
|---|---|---|---|---|---|---|---|---|
IBMOfficial | Cloud TPU / Nvidia H100 | 148 t/s | 160 ms (220 ms p90) | $0.08 | $0.2 | $0.008 | $0.11 | 99.99% |
| Groq LPU v2 (SRAM) | 622 t/s | 90 ms (120 ms p90) | $0.07 | $0.17 | $0.006 | $0.1 | 99.95% | |
| CS-3 Wafer-Scale Engine | 1850 t/s | 120 ms (160 ms p90) | $0.07 | $0.18 | — | $0.1 | 99.90% | |
| Nvidia H100 SXM5 | 222 t/s | 130 ms (180 ms p90) | $0.05 | $0.14 | $0.005 | $0.07 | 99.95% | |
| Nvidia H100 / L40S | 192 t/s | 140 ms (200 ms p90) | $0.04 | $0.1 | $0.004 | $0.06 | 99.90% | |
| Nvidia B200 / H100 | 281 t/s | 110 ms (150 ms p90) | $0.05 | $0.13 | $0.005 | $0.07 | 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.