OpenAIAPI Host Benchmarks
GPT-oss-20B — API Providers & Speed Matrix
Compare live throughput, TTFT latency distributions, prompt cache pricing, and hardware stacks across cloud hosting endpoints.
Fastest Provider
1125 t/s
Lowest Latency (TTFT)
0.09s P50
Best Price Value
$0.02 / $0.1
Blended Pricing Workload Calculator
Switch between standard conversational ratios, agentic caching loops, and heavy RAG workloads.
| Provider & Host | Hardware Stack | Output Speed (TPS) | Latency TTFT (P50) | Input Price ($/1M) | Output Price ($/1M) | Cache Read ($/1M) | Blended Price ($/1M) | SLA |
|---|---|---|---|---|---|---|---|---|
OpenAIOfficial | Cloud TPU / Nvidia H100 | 90 t/s | 0.8s (1.12s p90) | $0.05 | $0.2 | $0.005 | $0.09 | 99.99% |
Groq Cloud | Groq LPU v2 (SRAM) | 378 t/s | 0.09s (0.12s p90) | $0.04 | $0.17 | $0.004 | $0.07 | 99.95% |
Cerebras Cloud | CS-3 Wafer-Scale Engine | 1125 t/s | 0.12s (0.16s p90) | $0.05 | $0.18 | — | $0.08 | 99.90% |
Together AI | Nvidia H100 SXM5 | 135 t/s | 0.64s (0.88s p90) | $0.03 | $0.14 | $0.003 | $0.06 | 99.95% |
DeepInfra | Nvidia H100 / L40S | 117 t/s | 0.72s (1s p90) | $0.02 | $0.1 | $0.003 | $0.04 | 99.90% |
Fireworks AI | Nvidia B200 / H100 | 171 t/s | 0.56s (0.76s p90) | $0.03 | $0.13 | $0.003 | $0.06 | 99.99% |