GoogleAPI Host Benchmarks
Gemini 2.5 Flash — API Providers & Speed Matrix
Compare live throughput, TTFT latency distributions, prompt cache pricing, and hardware stacks across cloud hosting endpoints.
Fastest Provider
2738 t/s
Lowest Latency (TTFT)
0.09s P50
Best Price Value
$0.14 / $1.25
Blended Pricing Workload Calculator
Switch between standard conversational ratios, agentic caching loops, and heavy RAG workloads.
| Provider & Host | Hardware Stack | Output Speed (TPS) | Latency TTFT (P50) | Input Price ($/1M) | Output Price ($/1M) | Cache Read ($/1M) | Blended Price ($/1M) | SLA |
|---|---|---|---|---|---|---|---|---|
GoogleOfficial | Cloud TPU / Nvidia H100 | 219 t/s | 0.44s (0.62s p90) | $0.3 | $2.5 | $0.03 | $0.85 | 99.99% |
Groq Cloud | Groq LPU v2 (SRAM) | 920 t/s | 0.09s (0.12s p90) | $0.26 | $2.13 | $0.024 | $0.73 | 99.95% |
Cerebras Cloud | CS-3 Wafer-Scale Engine | 2738 t/s | 0.12s (0.16s p90) | $0.27 | $2.25 | — | $0.77 | 99.90% |
Together AI | Nvidia H100 SXM5 | 329 t/s | 0.35s (0.48s p90) | $0.2 | $1.75 | $0.018 | $0.59 | 99.95% |
DeepInfra | Nvidia H100 / L40S | 285 t/s | 0.4s (0.55s p90) | $0.14 | $1.25 | $0.015 | $0.42 | 99.90% |
Fireworks AI | Nvidia B200 / H100 | 416 t/s | 0.31s (0.42s p90) | $0.18 | $1.63 | $0.018 | $0.54 | 99.99% |