GoogleAPI Host Benchmarks
Gemini 3.1 Pro — API Providers & Speed Matrix
Compare live throughput, TTFT latency distributions, prompt cache pricing, and hardware stacks across cloud hosting endpoints.
Fastest Provider
1700 t/s
Lowest Latency (TTFT)
0.09s P50
Best Price Value
$0.9 / $6
Blended Pricing Workload Calculator
Switch between standard conversational ratios, agentic caching loops, and heavy RAG workloads.
| Provider & Host | Hardware Stack | Output Speed (TPS) | Latency TTFT (P50) | Input Price ($/1M) | Output Price ($/1M) | Cache Read ($/1M) | Blended Price ($/1M) | SLA |
|---|---|---|---|---|---|---|---|---|
GoogleOfficial | Cloud TPU / Nvidia H100 | 136 t/s | 32.39s (45.35s p90) | $2 | $12 | $0.2 | $4.5 | 99.99% |
Groq Cloud | Groq LPU v2 (SRAM) | 571 t/s | 0.09s (0.12s p90) | $1.7 | $10.2 | $0.16 | $3.82 | 99.95% |
Cerebras Cloud | CS-3 Wafer-Scale Engine | 1700 t/s | 0.12s (0.16s p90) | $1.8 | $10.8 | — | $4.05 | 99.90% |
Together AI | Nvidia H100 SXM5 | 204 t/s | 25.91s (35.63s p90) | $1.3 | $8.4 | $0.12 | $3.08 | 99.95% |
DeepInfra | Nvidia H100 / L40S | 177 t/s | 29.15s (40.49s p90) | $0.9 | $6 | $0.1 | $2.17 | 99.90% |
Fireworks AI | Nvidia B200 / H100 | 258 t/s | 22.67s (30.77s p90) | $1.2 | $7.8 | $0.12 | $2.85 | 99.99% |