These are very big "if"s. This year TSMC did not deliver a major new process upgrade and Apple did not deliver major CPU architecture upgrade. You may end up with 5% efficiency improvement and 2% performance improvement (if any).
They look ok but we need to wait for thorough testing. Individual benchmarks offer limited perspective. For example, Geekbench is known to be sensitive to memory performance. If Apple improved memory speed this would help Geekbench but not other (less memory sensitive) benchmarks.