TencentAPI Host Benchmarks
Hunyuan Hy3 — API Providers & Speed Matrix
Compare live throughput, TTFT latency distributions, prompt cache pricing, and hardware stacks across cloud hosting endpoints.
Fastest Provider
863 t/s
Lowest Latency (TTFT)
0.09s P50
Best Price Value
$0.06 / $0.28
Blended Pricing Workload Calculator
Switch between standard conversational ratios, agentic caching loops, and heavy RAG workloads.
| Provider & Host | Hardware Stack | Output Speed (TPS) | Latency TTFT (P50) | Input Price ($/1M) | Output Price ($/1M) | Cache Read ($/1M) | Blended Price ($/1M) | SLA |
|---|---|---|---|---|---|---|---|---|
TencentOfficial | Cloud TPU / Nvidia H100 | 69 t/s | 2.76s (3.86s p90) | $0.14 | $0.56 | $0.014 | $0.25 | 99.99% |
Groq Cloud | Groq LPU v2 (SRAM) | 290 t/s | 0.09s (0.12s p90) | $0.12 | $0.48 | $0.011 | $0.21 | 99.95% |
Cerebras Cloud | CS-3 Wafer-Scale Engine | 863 t/s | 0.12s (0.16s p90) | $0.13 | $0.5 | — | $0.22 | 99.90% |
Together AI | Nvidia H100 SXM5 | 104 t/s | 2.21s (3.04s p90) | $0.09 | $0.39 | $0.008 | $0.17 | 99.95% |
DeepInfra | Nvidia H100 / L40S | 90 t/s | 2.48s (3.45s p90) | $0.06 | $0.28 | $0.007 | $0.12 | 99.90% |
Fireworks AI | Nvidia B200 / H100 | 131 t/s | 1.93s (2.62s p90) | $0.08 | $0.36 | $0.008 | $0.15 | 99.99% |