Hi, I run [onthebench.ai](https://onthebench.ai), an open benchmark that measures LLM-gateway overhead (latency, throughput, memory, streaming, protocol translation) on neutral hardware. LiteLLM appears on the board twice, the Python proxy and the Rust `/v1/messages` beta, each benchmarked separately. I want to make sure I'm testing both fairly.

How it works, so there are no surprises: - Every gateway runs on the **same rig, same mock upstream, same load, same CPU pinning**, no per-gateway special-casing. - Each entry is defined by a single file: [`gateways/litellm-python/gateway.sh`](https://github.com/GetBusbar/benchmarking/blob/main/gateways/litellm-python/gateway.sh) and [`gateways/litellm-rust/gateway.sh`](https://github.com/GetBusbar/benchmarking/blob/main/gateways/litellm-rust/gateway.sh). Those files are the whole story of how I configured them. - Every number regenerates from committed JSON; the [method is documented here](https://onthebench.ai/gateways/method) and the whole thing is open source and re-runnable.

My guiding rule is that **a failure is my bug until proven the gateway's**. If a cell shows red or a number looks off, I'd rather find out I configured LiteLLM w…