personally, I disregard and ignore cross-language benchmarks. Benchmarks are hard, benchmarks in managed languages are extra hard.

And then the person writing the code has to be an expert in all languages involved to make it even half fair, instead of just a sound-bite/click-bait.