On the Bench · AI tool benchmarks
Reproducible, honest benchmarks for the AI tooling you actually run.
Every entrant is measured the same way, on neutral hardware: same box, same mock upstream, same load, same CPU pin.
Every number regenerates from committed JSON. Fully open source, re-runnable by anyone.
Built and operated by the Busbar team, who are also one of the entrants. Same harness for everyone. A failure is our bug until proven the gateway's.
Published OpenAI → OpenAI measurements
Gateway overhead against a local mock. Independent runs; compare configurations before drawing ratios.
| Gateway | Added p99 (µs) | State | Measured at |
|---|---|---|---|
| agentgateway | 251 | Valid | |
| AISIX (api7) | 174 | Valid | |
| APISIX | 340 | Valid | |
| Bifrost | 1,057 | Valid | |
| Busbar | 89 | Valid | |
| GoModel | 565 | Valid | |
| Helicone | 274 | Valid | |
| Higress | 943 | Valid | |
| Kong | 410 | Valid | |
| LiteLLM · Python | 9,522 | Valid | |
| LiteLLM · Rust | 279 | Valid | |
| One-API | 589,609 | Valid | |
| Plano | 205,966 | Valid | |
| Portkey | 3,902 | Valid | |
| TensorZero | 41,966 | Valid |