Thinking MachinesAPI Host Benchmarks
Inkling — API Providers & Speed Matrix
Compare live throughput, TTFT latency distributions, prompt cache pricing, and hardware stacks across cloud hosting endpoints.
Fastest Provider
1013 t/s
Lowest Latency (TTFT)
0.09s P50
Best Price Value
$0.45 / $2.02
Blended Pricing Workload Calculator
Switch between standard conversational ratios, agentic caching loops, and heavy RAG workloads.
| Provider & Host | Hardware Stack | Output Speed (TPS) | Latency TTFT (P50) | Input Price ($/1M) | Output Price ($/1M) | Cache Read ($/1M) | Blended Price ($/1M) | SLA |
|---|---|---|---|---|---|---|---|---|
Thinking MachinesOfficial | Cloud TPU / Nvidia H100 | 81 t/s | 1.82s (2.55s p90) | $1 | $4.05 | $0.1 | $1.76 | 99.99% |
Groq Cloud | Groq LPU v2 (SRAM) | 340 t/s | 0.09s (0.12s p90) | $0.85 | $3.44 | $0.08 | $1.5 | 99.95% |
Cerebras Cloud | CS-3 Wafer-Scale Engine | 1013 t/s | 0.12s (0.16s p90) | $0.9 | $3.65 | — | $1.59 | 99.90% |
Together AI | Nvidia H100 SXM5 | 122 t/s | 1.46s (2s p90) | $0.65 | $2.83 | $0.06 | $1.2 | 99.95% |
DeepInfra | Nvidia H100 / L40S | 105 t/s | 1.64s (2.27s p90) | $0.45 | $2.02 | $0.05 | $0.84 | 99.90% |
Fireworks AI | Nvidia B200 / H100 | 154 t/s | 1.27s (1.73s p90) | $0.6 | $2.63 | $0.06 | $1.11 | 99.99% |