Apple Mac Studio M3 Ultra runs massive AI models in memory, challenging traditional multi-GPU setups with efficient ...

Instead, the latest top-end Mac Studio features the long-awaited M3 Ultra. Similar to the M1 ... of memory bandwidth running something like DeepSeek R1 at home is totally feasible, at 4-bit ...

YouTuber Dave Lee of Dave2D fame has demonstrated how Apple's new Mac Studio equipped with an M3 Ultra chip can efficiently run a huge version ...

TL;DR: Apple's new M3 Ultra processor in the Mac Studio features a 32-core CPU and 80-core GPU, excelling in DeepSeek R1 model performance. It supports up to 512GB of unified RAM, outperforming ...

The Mac Studio is a great system for running AI models like DeepSeek locally. That is, if you're prepared to pay for M3 Ultra and a lot of upgrades.

DeepSeek's free 685B-parameter AI model runs at 20 tokens/second on Apple's Mac Studio, outperforming Claude Sonnet while ...

Apple launched its new M3 Ultra chip inside of the new Mac Studio, with a new Mac Cluster involving 4 systems using around 33W at idle, less than a switch.

As Ben Thompson pointed out this week, an M3 Ultra Mac Studio with 512GB of RAM could run the four-bit quantized version of the state-of-the-art DeepSeek R1 model on device. Even at a price tag of ...

When asked to improve the code by hiding the log file, DeepSeek returned code meeting that aim and carried only one critical ...

A Mac Studio with an M3 Ultra inside should have enough oomph to run something like DeepSeek R1 locally, without having to connect to servers in China.

etc etc etc I have to wonder, is the M3 Ultra a way of differentiating the Studio ... this on our new Mac Pro. We used 16 of them connected by Thunderbolt 5 to train a 450B DeepSeek-code-based ...