Catalogue
Models
Each model is served by more than one contracted supplier where we have one, and every request goes to whichever is cheaper for that request's shape, failing over if one is down. You pay the price on this page.
0 models · prices per million tokens at the standard rate, no platform fee per request · uptime, latency and throughput from the last 24 hours of real traffic.
No models match
Clear the search or the filters.
Use any model with one base URL.
https://api.yunshurouter.com/v1 takes OpenAI and Anthropic request bodies. Add "models": ["deepseek-v4-pro", "deepseek-v4.1-flash"] to fall back automatically if the first is unavailable.