Moonbit developers are lying to you
The Moonbit team recently published a blog post claiming their language runs "30% faster than Rust" for FFT workloads. This is a lie by omission. They benchmarked against a deliberately crippled Rust implementation that no competent programmer would write.
- The Moonbit FFT benchmark used a crippled Rust baseline and used to claim their language was faster than Rust.
- My corrected Rust implementation is 3.2–3.4× faster than Moonbit on the same benchmark.
- In 5 minutes of prompting GPT-5, I produced a Rust version already 2.33× faster than Moonbit.
- Zero PRs merged or replied to by the team at time of writing. There are PRs fixing the Rust benchmark older than their tweet announcing Moonbit was faster than Rust.
- Moonbit devs are programming language developers that have marketed their language aggressively on the basis of performance for awhile now, they know better than this.
- Moonbit should retract or clearly amend their blog post with corrected Rust baseline results. Including the qualification that their benchmark is a naive Cooley-Tukey FFT benchmark and nothing else.