Apple M4 Max CPU transcribes audio twice as fast as the RTX A5000 GPU in user test — M4 Max pulls just 25W compared to the RTX A5000’s 190W

Apple’s all-new M4 silicon excels significantly in transcribing compared to other CPUs and GPUs, even discrete GPUs. Iniysa on X reports that Apple’s M4 Max accomplished an audio transcode with Whisper V3 Turbo in half the time of Nvidia’s Ampere-based RTX A5000 GPU while using nearly eight times less power.

The M4 Max purportedly transcribed an audio file using Whisper V3 Turbo in just 2:29 minutes with MLX while consuming 25 watts of power. The RTX A5000 graphics card purportedly ran the same test in 4:33 while consuming 190 watts of power. That is a 7.6x power consumption advantage in favor of the M4 Max.

It’s worth mentioning that Nvidia’s RTX A5000 GPU is a previous-generation part, so newer GPUs based on Ada Lovelace will most certainly perform better in this test. Regardless, the performance margin between the two is outstanding, especially considering the power consumption of both chips.

The RTX A5000 is an Ampere-based professional GPU featuring 8,192 CUDA cores, 24GB of GDDR6 memory, 768 GB/s of memory bandwidth, and a total board power rating of 230W. The M4 Max, by contrast, is an APU technically featuring 12 performance cores and four efficiency cores paired with a beefy media engine that supports all modern encoding and decoding standards, including AV1. The GPU has ten cores in total.

The main advantage of the M4 Max is its large number of encoders. Rather than having just one or two encoders, as many GPUs today do, the M4 Max features four total encoders, two regular video encode engines, and two Pro Res encode and decode engines.

Undoubtedly, Apple’s investment in heavy encoding performance helped fuel the M4 Max’s impressive transcoding win against the RTX A5000. More…