Those benchmarks should be done by someone independent from one of the tested frameworks though. Hard to trust the methodology, especially with Elm emerging as the "winner".