Which Macs Will Get an M6 Chip and Which Won't?

[MacRumors]

Apple announced new Mac mini machines this week, and the lower-end models include the first M6 chip. The M6 era will be a little complicated, because few Macs are going to get it. We won't have M6 Pro or M6 Max chips, which shakes up the Mac lineup for the next couple of years.

Subscribe to the MacRumors YouTube channel for more videos.​

Devices With M6

The Mac mini is the first Mac with an M6, but the chip is going to come to a couple of other products in the lineup.

• MacBook Pro - The 14-inch entry-level MacBook Pro could be updated with an M6 chip in October 2026.
• iMac - The 24-inch iMac is also expected to get the M6 chip, likely around the time the MacBook Pro is updated.
• MacBook Air - The MacBook Air could get the M6 chip in early 2027.
• iPad Pro - The next iPad Pro could get the M6 chip.
• iPad Air - A future version of the iPad Air is likely to adopt the M6. Apple usually keeps the iPad Air processor one generation behind the iPad Pro.

M6 Pro and M6 Max

Apple isn't going to make M6 Pro or M6 Max chips, according to Bloomberg, so the M6 will be the first standalone Apple silicon chip that won't have more powerful variants.

Instead of manufacturing M6 Pro and M6 Max chips, Apple is speeding up its chip development timeline to bring the M7 Pro and M7 Max to market more quickly. The M7 Pro and M7 Max will have technologies that support on-device AI and GPU-intensive software, so they better meet growing demand for processors with those capabilities.
M5 Makes Up the Gap

Macs that would have used an M6 Pro or M6 Max chip are going to be stuck with the M5 Pro and M5 Max until the M7 family is ready. That means the high-end OLED "MacBook Ultra" we're expecting in late 2026 or early 2027 is going to use the same M5 Pro and M5 Max chips that are in the current MacBook Pro models. It's also possible it will get the M5 Ultra.

The new Mac mini also got the M5 Pro instead of an M6 Pro chip. The new Mac Studio announced alongside the Mac mini uses M5 Max and M5 Ultra chips.
Short Life Span

The M6 isn't going to be around for long, because Apple plans to launch the M7 in the first half of 2027. The M7 Pro and M7 Max will follow at the end of the year.
M6 Features

The M6 is Apple's first chip built on a new 2-nanometer process that increases transistor density on a smaller die. The process change improves performance and efficiency.

It has a 12-core CPU with what Apple says is the world's fastest CPU core, a larger 12-core GPU with Neural Accelerators for each core, a new Dual 16-core Neural Engine, and up to 170GB/s memory bandwidth.

The GPU has updated shader core architecture, Dynamic Caching, hardware-accelerated ray tracing, and 50 percent higher geometry rates for complex graphics.

Article Link: Which Macs Will Get an M6 Chip and Which Won't?

 

[Nyx1]

Still not using WMCM packaging and lpddr6 wonder how A20 pro come out are they use new dram lpddr6 with WMCM packaging

 

[minhrom]

give me this Apple: iMac 27-inch, M6 , 16Gb, 256GB for $1199.

 

[Corefile]

I won't buy another M-series machine until M9 in 2029. Sorry Apple but you've blown it.

 

[jz0309]

well, none of my devices will get an M6 (replacement), 17PM going strong as is my Neo, Andy M1Max Studio as well, the Studio will continue to get AC+ for a few more years, don't need a replacement ...

 

[applesith]

Perhaps there are technical reasons, but why not just wait until there is a full family of chips and make that the M6 family instead of splitting it across M6 and M7?

 

[ikramerica]

I could see this being a pattern where the odds are pro/max/ultra and the evens are just chips.

“Consumer” models updated in fall each year with M chips the same generation as the iPhone A chips, and “professional” models updated less often but spring/summer with odd number M chips.

 

[svish]

Think the next iPad Pro could get M7. Waiting to see the new iMac. Hopefully the colors will continue to be bright.

 

[MacMorrison]

Still no native AV1 encode... I believe the M6 is a transitional product.

The TSMC 2nm process is new, but the performance improvements are incremental in this generation.

 

[JPack]

Perhaps there are technical reasons, but why not just wait until there is a full family of chips and make that the M6 family instead of splitting it across M6 and M7?

Probably because Apple realized half way into developing M6 that it wouldn't be able to handle future AI workloads. They cut the cord on M6 Pro/Max variants and went on to M7. But there's no reason to dump all M6 R&D.

 

[Siliconguy]

Probably because Apple realized half way into developing M6 that it wouldn't be able to handle future AI workloads. They cut the cord on M6 Pro/Max variants and went on to M7. But there's no reason to dump all M6 R&D.

"Future AI workloads" imply high power consumption. No problem on a desktop and tolerable on a Pro laptop. On an iPad or MacBook Air with limited battery capacity and passive cooling it is a problem. So M-Even# battery life first, M-Odd# compute/GPU first makes a great deal of sense.

Engineering is always about compromises.

 

[HVDynamo]

Probably because Apple realized half way into developing M6 that it wouldn't be able to handle future AI workloads. They cut the cord on M6 Pro/Max variants and went on to M7. But there's no reason to dump all M6 R&D.

Yeah, it's most likely they wanted to make that pivot and this was the best way to do it without wasting too much investment.

 

[Bug-Creator]

Well they kinda bought the 2nm supply well in advance, they have a chip design that works for the "low" range, so what are they supposed to do?

Pay a penalty for not using the 2nm supply while every other node is in short supply?

 

[jole]

give me this Apple: iMac 27-inch, M6 , 16Gb, 256GB for $1199.

And, here it is for only¹ $1599

¹ BYOC (Bring Your Own Compute) — add any amount of processing power, memory, and storage you choose by attaching a Mac Mini or Mac Studio. Sold separately.

 

[71Games]

However I look at it, it's really expensive tech. Small xcode builds still work on base macbook air m1. What really hurts is 13 inch screen. I hate xcode on 13 inch screen. I run away from it as quickly as I can. Glad Windows machines exist & more affordable price tags.

I'd love 16 inch pro but it's not going to happen. Second hand older pro is not the way due to shorter and shorter software support lifespan, and still relatively high price tag.

So I keep on reading about m2, m3, m4, m5, m6-7... Cool stuff and that's about it. 🙂

 

[Elwe]

Probably because Apple realized half way into developing M6 that it wouldn't be able to handle future AI workloads. They cut the cord on M6 Pro/Max variants and went on to M7. But there's no reason to dump all M6 R&D.

This is only trade press speculation, but in addition to what you wrote, there is educated guessing that TSMCs yields (and so costs) for the second generation of its 2nm chips is much better. They (and Apple) may have realized this earlier than in normal cycles, and so Apple decided it is worth it to move dependent products cycles forward.

Unfortunately, we will not know for sure for a while. At some point, we’ll get a clear idea of what M7 actually is (features, performance, power envelope, etc) and then we can judge. Maybe a next generation of neural accelerator; maybe much higher throughput memory and capacities; maybe enhanced encoding/decoding functionality. Not many people guessed Apple would introduce a second NPU into a base chip (i.e., M6), so maybe the M7 Pro will get more than two.

 

[Vincent Hanna]

That means the high-end OLED "MacBook Ultra" we're expecting in late 2026 or early 2027 is going to use the same M5 Pro and M5 Max chips that are in the current MacBook Pro models. It's also possible it will get the M5 Ultra.

No, it is NOT possible. No MacBook will ever get an Ultra chip.

 

[boswald]

If the M6 is a short-lived processor, does that mean the support window will be shorter as well?

(I'm sorry if that's a dumb question.)

 

[DanneP]

give me this Apple: iMac 27-inch, M6 , 16Gb, 256GB for $1199.

That would be cool, but Apple is charging $1599 for just the 27-inch part so it seems a bit unlikely.

 

[JPack]

This is only trade press speculation, but in addition to what you wrote, there is educated guessing that TSMCs yields (and so costs) for the second generation of its 2nm chips is much better. They (and Apple) may have realized this earlier than in normal cycles, and so Apple decided it is worth it to move dependent products cycles forward.

Unfortunately, we will not know for sure for a while. At some point, we’ll get a clear idea of what M7 actually is (features, performance, power envelope, etc) and then we can judge. Maybe a next generation of neural accelerator; maybe much higher throughput memory and capacities; maybe enhanced encoding/decoding functionality. Not many people guessed Apple would introduce a second NPU into a base chip (i.e., M6), so maybe the M7 Pro will get more than two.

There's not much to support the yield theory though. TSMC has a couple dozen customer tape outs for N2. And N2X and N2P are incremental refinements according to TSMC. Apple launched M3 Pro/Max/Ultra with the infamous N3B. None of this supports the idea that M6 Pro/Max was dropped because of TSMC performance.

With the rapid advances in AI and how Apple was caught flatfooted on the software side, it's no surprise on the silicon side, Apple would need a right hand turn as well.

 

[JMalone]

"Future AI workloads" imply high power consumption. No problem on a desktop and tolerable on a Pro laptop. On an iPad or MacBook Air with limited battery capacity and passive cooling it is a problem. So M-Even# battery life first, M-Odd# compute/GPU first makes a great deal of sense.

Engineering is always about compromises.

The typical user of local AI models gets about 1 hour a day of generation time between prompts. So for local models daily energy consumption is no more intensive than gaming.

Cloud models are even more efficient per user because of their tokens per watt performance, but that’s another subject. The public has been brainwashed by anti developmental trouble makers to believe AI literally drinks all the water in the world and burns down planets.

 

[Pezimak]

If you are vaguely interested in Ai but want a computer that will last, and you want a Mac Studio, is it worth hanging on of the M7 with the new 2nm architecture or is the M5 Max more then enough? Assuming the Studio gets the M7 Max IF these rumours are true of course.