The performance of _.each should be very close to that of the native Array#forEach because it delegates to the native method when it's available: https://github.com/jashkenas/underscore/blob/6db5b93eac3eacd...