Apple's M6 MacBook Pro Starts at $1,999 and Is Mostly a Chip Bump. If You Own an M5, Keep Your Money.
Apple announced a refreshed 14-inch MacBook Pro with the M6 chip on October 27. It starts at $1,999, pre-orders open October 28, and it ships November 6. By Apple's own numbers it is up to 30% faster in graphics and up to 25% faster on AI workloads than the M5 model, at the same starting price. That is a good deal on a new laptop and a weak reason to replace one you already own.
What changed
Per MacRumors, the M6 has a 12-core CPU (2 super cores, 4 performance cores and 6 efficiency cores) where the M5 had 10, and a 12-core GPU where the M5 had 10. Memory bandwidth goes from 153GB/s to up to 170GB/s, though MacRumors says that bump applies to the 24GB and 32GB configurations only. The wireless chip moves to Wi-Fi 7 and Bluetooth 6, and the machine can drive two external 4K displays at up to 165Hz. TechCrunch lists the same hardware, with the usual ports, a 12-megapixel camera and a claimed 24 hours of battery life.
The base model has 16GB of memory and 1TB of storage, with configurations up to 32GB and 4TB. MacRumors also notes Apple dropped the 512GB storage tier in March. A 24-inch M6 iMac was announced alongside it, starting at $1,499.
MacRumors summarizes the rest as the same as the M5 model. I would read this as a chip refresh, not a redesign.
What the AI claim does and does not mean
Apple says up to 25% faster AI workloads than M5, and up to 13 times the LLM prompt-processing speed of the M1 MacBook Pro. Those are vendor numbers for the best case, measured on Apple's chosen workloads. I have not seen independent benchmarks yet, and I would wait for them before believing any percentage on a product page.
For anyone running local models, I would look at the memory ceiling before the chip. A 32GB maximum caps how large a model you can hold, and a model that does not fit is slower than any chip improvement can fix. The bandwidth figure only improves on the bigger memory configs, which means the cheapest M6 does not get the headline memory upgrade. If local inference is a real part of your workflow, price out the 32GB config, not the $1,999 one.
A buy rule for a one-person shop
A laptop is a business tool with a payback period, so I would ask three questions.
Is your current machine costing you paid hours? If builds, test runs or local model loads make you wait more than a few minutes a day, a 25% speedup might be worth real money over a year. If your waiting time is mostly in the cloud, nothing here helps.
Is your current machine failing in a way a chip will not fix? Battery wear, a bad display, or a machine that is out of its support window are legitimate reasons to replace it, and the M6 is a fine thing to replace it with.
Are you buying on a deadline or a tax year? If you want the expense this year, buying before December is a bookkeeping decision. That is a different question from whether the laptop is faster.
If you own an M5 MacBook Pro and none of those apply, I would skip it. If you are on an older machine and one of them does, the M6 is a reasonable thing to buy.
What I could not confirm
Pre-launch coverage expected an OLED, touchscreen MacBook Pro to be announced around the same time. I could not confirm from the pages I read that it launched, and MacRumors' own pre-order page does not describe one. If you are considering a high-end MacBook Pro, check Apple's site before ordering, because a new design might change the math for the top of the lineup. I am only making a claim about the 14-inch M6 here.
What I would actually do
Hold off if you own an M5. If you are replacing something older, decide on memory first (24GB is my floor if you touch local models at all), and wait for independent benchmarks after November 6 before spending extra on a bigger chip. Treat Apple's launch percentages as the ceiling, not the forecast.
Author
Lukas
@lukcombinator