If (Big if) the provided graphs on apple's M1 Pro & M1 Max article are to be believed you're getting 90% the performance of a RTX 3080 Mobile (A 3070 with a clockspeed cut) but at a quarter of the power draw. This seems to be quite a development.
For all the other claims on GPU performance, they call out that they are comparing to the integrated graphics, and they name the GPU specifically. With this one claim, they name a computer, but not the GPU. "Compared to a computer that has a laptop 3080 in it" is not the same thing as "compared to a 3080."
And they don't define "performance". Is it fill rate? Is it TFLOPs? Is it memory bandwidth? The M1 Max has 64 GPU cores. That sounds like a lot! It's not! The NVidia Laptop 3080 has 6,144 cores. That's NINETY SIX TIMES as many cores. I'm sorry, but supposedly achieving 90% of the "performance" of a GPU that has 2 orders of magnitude more cores is just not in the realm of believable. They are being intentionally evasive here.
The GPU performance of my M1 Mac Mini is roughly comparable to that of my Pixel 5 and Quest 2--both of which are running the Qualcomm Adreno 650 GPU.