Apple Launches M6 MacBook Pro and iMac With a Bigger On-Device AI Pitch

Apple announced a 14-inch MacBook Pro and a new iMac powered by its M6 chip on Tuesday, putting local AI processing at the center of another Mac refresh. The notebook starts at $1,999 in the United States and the desktop at $1,499. Preorders open October 28, with customer deliveries and retail availability beginning November 6.

The launches give Apple two different ways to sell the same broad proposition: more computing work performed on the device rather than sent to a remote server. That can matter for responsiveness, offline operation and the handling of sensitive material. It also makes software compatibility and memory capacity important parts of the buying decision, beyond a processor's headline performance number.

Reuters confirmed the announcement and release schedule. These are announced products with published specifications, separate from the speculation surrounding Apple's October 28 event. Our earlier event preview covered what was expected; Tuesday's releases now establish which changes are official without confirming every other rumored device or feature.

The laptop and desktop use different benchmarks

Apple says the M6 MacBook Pro offers up to 30% faster graphics and up to 25% faster AI workloads than the M5 model. Its claim of up to five times faster on-device AI for the iMac uses the older M4 as the comparison. Those percentages cannot be placed side by side as though they measure the relative speed of the two new computers.

The distinction is practical for existing owners. A result measured against an older generation can show a much larger improvement than one measured against the immediate predecessor. Specific tasks can also stress different parts of the system. Prompt processing, image generation and a long video export do not necessarily respond equally to extra CPU cores, graphics resources or dedicated AI hardware.

Apple's figures are manufacturer test claims, not independent TTN measurements. An “up to” result describes a selected test outcome rather than the improvement every application will deliver. Anyone comparing upgrade options needs the relevant workload, configuration and baseline. A large improvement in one benchmark does not establish that ordinary document editing or browsing will feel proportionately faster.

Memory can decide which local models are useful

The notebook's official configuration includes a 12-core CPU, a 12-core GPU and two 16-core Neural Engines. It starts with 16GB of unified memory and can be configured with up to 32GB. MacRumors reports that the base notebook includes 1TB of storage. Memory and storage perform different jobs: more SSD capacity does not replace the memory required to keep an AI model and its working data available.

Our analysis of AI returns describes why usable workflow improvements matter more than a demonstration alone. A local model that loads successfully still needs to produce sufficiently accurate results at an acceptable speed. Editing, checking and integrating its output remain part of the total work, particularly for professional tasks where errors have a cost.

Local processing can reduce the amount of information sent elsewhere, but it is not a blanket description of everything an application does. A program may combine device-based features with online services, account synchronization or cloud inference. The data-security implications depend on the actual workflow and permissions, rather than simply on whether the computer contains a faster AI accelerator.

The iMac offers a different upgrade calculation

The new iMac retains a 24-inch 4.5K display and seven color options. Apple lists a $1,499 version with a 10-core CPU and GPU and a $1,699 version with 12 cores of each. Both product families add Wi-Fi 7 and Bluetooth 6. Those wireless specifications are capabilities, not guarantees of a faster internet connection without suitable networking equipment.

A desktop buyer gets an integrated display and a fixed workstation; a notebook buyer pays for portability as well as computing capacity. That difference can outweigh a benchmark gap. Tuesday's announcements supply concrete configurations and dates for that comparison. Independent application testing, once the machines are available, will provide the missing evidence about how consistently Apple's claimed gains appear in everyday professional use.