Z.aiAPI Host Benchmarks
GLM-5.2 — API Providers & Speed Matrix
Compare live throughput, TTFT latency distributions, prompt cache pricing, and hardware stacks across cloud hosting endpoints.
Fastest Provider
2050 t/s
Lowest Latency (TTFT)
0.09s P50
Best Price Value
$0.61 / $2.15
Blended Pricing Workload Calculator
Switch between standard conversational ratios, agentic caching loops, and heavy RAG workloads.
| Provider & Host | Hardware Stack | Output Speed (TPS) | Latency TTFT (P50) | Input Price ($/1M) | Output Price ($/1M) | Cache Read ($/1M) | Blended Price ($/1M) | SLA |
|---|---|---|---|---|---|---|---|---|
Z.aiOfficial | Cloud TPU / Nvidia H100 | 164 t/s | 1.51s (2.11s p90) | $1.35 | $4.29 | $0.135 | $2.08 | 99.99% |
Groq Cloud | Groq LPU v2 (SRAM) | 689 t/s | 0.09s (0.12s p90) | $1.15 | $3.65 | $0.108 | $1.77 | 99.95% |
Cerebras Cloud | CS-3 Wafer-Scale Engine | 2050 t/s | 0.12s (0.16s p90) | $1.22 | $3.86 | — | $1.88 | 99.90% |
Together AI | Nvidia H100 SXM5 | 246 t/s | 1.21s (1.66s p90) | $0.88 | $3 | $0.081 | $1.41 | 99.95% |
DeepInfra | Nvidia H100 / L40S | 213 t/s | 1.36s (1.89s p90) | $0.61 | $2.15 | $0.068 | $0.99 | 99.90% |
Fireworks AI | Nvidia B200 / H100 | 312 t/s | 1.06s (1.43s p90) | $0.81 | $2.79 | $0.081 | $1.31 | 99.99% |