| Route | p50 latency | p95 latency | Throughput | Success rate | ShareShare of traffic |
|---|---|---|---|---|---|
| Alibaba DashScope — China | 7.49 s | 33.57 s | 243 tok/s | 100.0% | 49% |
| DeepSeek direct | 1.67 s | 13.64 s | 99 tok/s | 100.0% | 25% |
| InferHub | 5.55 s | 33.57 s | 134 tok/s | 99.5% | 22% |
| Venice AI direct | 916 ms | 7.49 s | 99 tok/s | 100.0% | 2% |
| Dolphin | 45.32 s | 150.53 s | 54 tok/s | 60.1% | 2% |
| OpenRouter | 3.04 s | 24.87 s | 329 tok/s | 99.9% | 0% |
| Ollama Cloud direct | 2.25 s | 24.87 s | 243 tok/s | 99.3% | 0% |
| NeuralWatt | 4.11 s | 24.87 s | 180 tok/s | 56.2% | 0% |
Any OpenAI-compatible client works. Add /min30 to the path to only buy at 30% off or better.
Read the quickstart ↗from openai import OpenAI client = OpenAI( api_key="inf_your_key_here", base_url="$SURPLUS_BASE_URL" ) response = client.chat.completions.create( model="deepseek-v4.1-flash", messages=[{"role": "user", "content": "Hello!"}] )
SURPLUS_BASE_URL
https://api.surplusintelligence.ai/v1