| Route | p50 latency | p95 latency | Throughput | Success rate | ShareShare of traffic |
|---|---|---|---|---|---|
| Alibaba DashScope — China | 10.11 s | 33.57 s | 30 tok/s | 97.9% | 85% |
| BytePlus ModelArk | 13.64 s | 111.50 s | 40 tok/s | 99.6% | 6% |
| DeepSeek direct | 18.42 s | 82.59 s | 54 tok/s | 99.9% | 6% |
| InferHub | 13.64 s | 111.50 s | 54 tok/s | 87.4% | 1% |
| Uncensored AI direct | 10.11 s | 33.57 s | 73 tok/s | 100.0% | 1% |
| Venice AI direct | 45.32 s | 203.21 s | 30 tok/s | 99.6% | 1% |
| OpenRouter | 7.49 s | 61.18 s | - | 3.6% | 0% |
| Mordiem | 82.59 s | - | 22 tok/s | 74.3% | 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-pro", messages=[{"role": "user", "content": "Hello!"}] )
SURPLUS_BASE_URL
https://api.surplusintelligence.ai/v1