OpenAI deploys Pando Git backend, reports nearly 80% lower median push latencyMachine translation

Pando now serves all development in OpenAI's monorepo, while engineers and Codex agents continue using git push and git fetch as before. OpenAI reports nearly 80% lower median push latency; the figure has no independent benchmark yet, and the backend is not open source.