OpenAI/Codex relay selection should consider model availability, context continuity, tool-call stability, and cost control.
This ranking is based on 10 recent RelayRank monitoring samples, comparing availability, latency, current status, listing age, and traceable trends.
Current top providers
YunDou: 100% availability, 2011 ms average latency, current status up.
lingxicode: 99.8% availability, 2334 ms average latency, current status up.
SudoCode: 99.1% availability, 2249 ms average latency, current status up.
toproutercn: 98.9% availability, 2056 ms average latency, current status up.
aimz: 98.9% availability, 2391 ms average latency, current status up.
How to read it
This is not a top-up recommendation or a permanent verdict. Use it as a first screening layer before running small real-task tests.
Claude Code, Codex, Gemini, and Agent workflows should put sustained availability and status volatility ahead of one-off latency.