benchmark.rb

require 'benchmark/ips'

Benchmark.ips do |x|
  array = [[1, 2, 3], [4, 5, 6], [7, 8, 9]]

  x.report('transpose') { array.transpose.product(array.transpose).flatten(1) }
  x.report('proc') { :product.to_proc.call(*array) }
  x.report('reduce') { array.reduce(&:product).map(&:flatten) }
  x.report('then') { array.then { |head, *tail| head.product *tail } }
  x.report('slices') { array[0].product(array[1..-1]) }
  x.report('endless first slices') { array.first.product(array[1..]) }
  x.report('endless slices') { array[0].product(array[1..]) }

  x.compare!
end

cruby

Calculating -------------------------------------
           transpose    430.108k (± 2.9%) i/s -      2.151M in   5.004655s
                proc    485.482k (± 2.8%) i/s -      2.441M in   5.031551s
              reduce     61.228k (± 2.6%) i/s -    309.320k in   5.055422s
                then    466.964k (± 6.8%) i/s -      2.350M in   5.058004s
              slices      1.812M (± 9.4%) i/s -      9.033M in   5.033413s
endless first slices      1.620M (± 2.4%) i/s -      8.197M in   5.061826s
      endless slices      1.679M (± 3.0%) i/s -      8.490M in   5.060451s

Comparison:
              slices:  1812214.4 i/s
      endless slices:  1679289.2 i/s - same-ish: difference falls within error
endless first slices:  1620389.7 i/s - same-ish: difference falls within error
                proc:   485481.5 i/s - 3.73x  slower
                then:   466964.2 i/s - 3.88x  slower
           transpose:   430107.5 i/s - 4.21x  slower
              reduce:    61227.8 i/s - 29.60x  slower

jruby

Calculating -------------------------------------
           transpose    511.189k (± 4.8%) i/s -      2.568M in   5.036421s
                proc    737.483k (±10.0%) i/s -      3.678M in   5.046783s
              reduce     79.450k (± 6.2%) i/s -    396.312k in   5.009035s
          yield_self    747.791k (± 5.3%) i/s -      3.746M in   5.024160s
              slices      3.796M (± 8.8%) i/s -     18.796M in   4.997241s

Comparison:
              slices:  3796325.8 i/s
          yield_self:   747790.9 i/s - 5.08x  slower
                proc:   737483.2 i/s - 5.15x  slower
           transpose:   511189.2 i/s - 7.43x  slower
              reduce:    79450.0 i/s - 47.78x  slower

truffleruby

Calculating -------------------------------------
           transpose    378.467k (±45.3%) i/s -    701.387k in   5.049225s
                proc    278.442k (±22.8%) i/s -    555.768k in   5.202504s
              reduce     93.156k (±30.7%) i/s -    127.890k in   5.043050s
                then     58.353k (±74.0%) i/s -     65.100k in   5.014233s
              slices    899.148k (±24.6%) i/s -      1.298M in   4.996822s

Comparison:
              slices:   899148.2 i/s
           transpose:   378466.7 i/s - 2.38x  slower
                proc:   278441.5 i/s - 3.23x  slower
              reduce:    93156.2 i/s - 9.65x  slower
                then:    58352.6 i/s - 15.41x  slower