Yeah, but they still write regarding plain M6: "M6 also introduces a Dual 16-core Neural Engine, providing up to 2x the peak compute over previous generations

Yeah, but they still write regarding plain M6: "M6 also introduces a Dual 16-core Neural Engine, providing up to 2x the peak compute over previous generations to make on-device AI workflows run even faster" .