OpenAIAPI Host Benchmarks
GPT-5.6 Sol — API Providers & Speed Matrix
Compare live throughput, TTFT latency distributions, prompt cache pricing, and hardware stacks across cloud hosting endpoints.
Fastest Provider
900 t/s
Lowest Latency (TTFT)
0.09s P50
Best Price Value
$2.25 / $15
Blended Pricing Workload Calculator
Switch between standard conversational ratios, agentic caching loops, and heavy RAG workloads.
| Provider & Host | Hardware Stack | Output Speed (TPS) | Latency TTFT (P50) | Input Price ($/1M) | Output Price ($/1M) | Cache Read ($/1M) | Blended Price ($/1M) | SLA |
|---|---|---|---|---|---|---|---|---|
OpenAIOfficial | Cloud TPU / Nvidia H100 | 72 t/s | 137.75s (192.85s p90) | $5 | $30 | $0.5 | $11.25 | 99.99% |
Groq Cloud | Groq LPU v2 (SRAM) | 302 t/s | 0.09s (0.12s p90) | $4.25 | $25.5 | $0.4 | $9.56 | 99.95% |
Cerebras Cloud | CS-3 Wafer-Scale Engine | 900 t/s | 0.12s (0.16s p90) | $4.5 | $27 | — | $10.13 | 99.90% |
Together AI | Nvidia H100 SXM5 | 108 t/s | 110.2s (151.53s p90) | $3.25 | $21 | $0.3 | $7.69 | 99.95% |
DeepInfra | Nvidia H100 / L40S | 94 t/s | 123.98s (172.19s p90) | $2.25 | $15 | $0.25 | $5.44 | 99.90% |
Fireworks AI | Nvidia B200 / H100 | 137 t/s | 96.42s (130.86s p90) | $3 | $19.5 | $0.3 | $7.13 | 99.99% |