M5 Max compile times against M3 Max — 40 real projects, thermals logged
Same Xcode and Rust workloads on both machines, ambient held at 22C, full spreadsheet in the comments.
Short version: the gap is much bigger on sustained builds than on the single-run numbers everyone quotes. Cold-start benchmarks flatter the M5 by roughly 12%.
2413
3
Top
First benchmark post in months that actually logged ambient temperature. Thank you.
Does the gap hold on the base M5, or is this a Max-only story?
Base M5 numbers are in the sheet, tab two. Gap is smaller but the same shape.