xAIAPI Host Benchmarks
Grok 4.3 — API Providers & Speed Matrix
Compare live throughput, TTFT latency distributions, prompt cache pricing, and hardware stacks across cloud hosting endpoints.
Fastest Provider
1613 t/s
Lowest Latency (TTFT)
0.09s P50
Best Price Value
$0.56 / $1.25
Blended Pricing Workload Calculator
Switch between standard conversational ratios, agentic caching loops, and heavy RAG workloads.
| Provider & Host | Hardware Stack | Output Speed (TPS) | Latency TTFT (P50) | Input Price ($/1M) | Output Price ($/1M) | Cache Read ($/1M) | Blended Price ($/1M) | SLA |
|---|---|---|---|---|---|---|---|---|
xAIOfficial | Cloud TPU / Nvidia H100 | 129 t/s | 0.73s (1.02s p90) | $1.25 | $2.5 | $0.125 | $1.56 | 99.99% |
Groq Cloud | Groq LPU v2 (SRAM) | 542 t/s | 0.09s (0.12s p90) | $1.06 | $2.13 | $0.1 | $1.33 | 99.95% |
Cerebras Cloud | CS-3 Wafer-Scale Engine | 1613 t/s | 0.12s (0.16s p90) | $1.13 | $2.25 | — | $1.41 | 99.90% |
Together AI | Nvidia H100 SXM5 | 194 t/s | 0.58s (0.8s p90) | $0.81 | $1.75 | $0.075 | $1.04 | 99.95% |
DeepInfra | Nvidia H100 / L40S | 168 t/s | 0.66s (0.91s p90) | $0.56 | $1.25 | $0.063 | $0.73 | 99.90% |
Fireworks AI | Nvidia B200 / H100 | 245 t/s | 0.51s (0.69s p90) | $0.75 | $1.63 | $0.075 | $0.97 | 99.99% |