Apple M5 Ultra with 512GB Memory Runs Massive Open-Weight Models Locally
MaySaki2 · reddit · 2026-08-26
The highlight of Apple's new M5 Ultra is its up to 512GB of unified memory with 1.2TB/s bandwidth.
This allows a single Mac Studio to load and run massive models like DeepSeek R1 671B and Kimi K2.6 1T locally at usable quantizations. These are models typically associated with racks of GPUs, now fitting into a compact desktop machine.
- Deployment Shift: For researchers and developers, powerful AI is accessible without building complex multi-GPU systems, keeping data private and avoiding per-token costs.
- Market Split: Local AI hardware is bifurcating into affordable systems for small models and high-memory workstations bringing server-class models to a single box.
- Comparison: In contrast, the M6 Mac mini, despite faster AI hardware, is capped at 32GB, keeping it in the compact-model category.
The post includes a model compatibility list for the M5 Ultra and a guide for running LLMs on Apple Silicon.
More from Embodied
- Volvo and NVIDIA expand partnership to bring safe automated driving to next-gen cars — nvidia · 2026-10-16
- Tevel's robotic system harvests fruit around the clock using computer vision — LexiLove · 2026-10-16
- SHOES: 10 minutes of real-robot RL lets quadrupeds walk in shoes, CoRL 2026 paper — kscottz · 2026-10-16
- Tesla Cybercab robotaxis logged zero crashes in first two weeks of public rides, NHTSA data shows — elonmusk · 2026-10-16
- Robotics LLMs tested on 16 tasks: GPT-6 Astra hits 17.2% success vs Fable's 9.1% — pabbeel · 2026-10-16
- WEDGE paper generates training environments conditioned on navigation agent failures — _akhaliq · 2026-10-16