Best Local LLMs for the Mac Studio M3 Ultra (512 GB)

The best local LLM for a Mac Studio M3 Ultra (512 GB) is MiMo-V2.6-Flash at 59 tok/s (community run). Its 512 GB fits 122 of the 130 models in the LLMCheck index, up to 1.05T-class. For everyday chat and coding, MiMo-V2.6-Flash is the sweet spot.

Unified memory
512
GB
Mem. bandwidth
819
GB/s
Models that fit
122
of 130
Top speed
estimated334
tok/s

Considering this Mac?
Mac Studio M5 Ultra (512GB): build-to-order from Apple · Mac Studio M3 Ultra on Amazon Renewed → · compare every Mac for local AI →

Top 3 picks for the Mac Studio M3 Ultra (512 GB)

⭐ Best overall
309B MoE · MIT · cap 49/50 · agent-ready 64K
community59 tok/s
⚡ Fastest
1.73B · LFM Open · cap 17/50 · agent-ready 64K
estimated334 tok/s
🧠 Most capable
552B MoE · MIT · cap 50/50 · agent-ready 64K raise GPU limit (macOS 15)
On macOS 15, raise the GPU limit first: sudo sysctl iogpu.wired_limit_mb=507904
community20 tok/s

Every model ranked for a Mac Studio M3 Ultra (512 GB)

Ranked by LLMCheck suitability (capability balanced against speed on the M3 Ultra); the top pick is the best-ranked model that reaches 15 tok/s here, has a standard Mac runtime and fits the GPU memory limit (raised if need be), and slower models are tagged slow. Click a model for its full figures and setup. Every speed carries its provenance mark: most are index estimates from the memory-bandwidth model, a filled square is a published figure and a hollow circle a community run — submit a real run →

#ModelSizeLicenseSpeed64K contextCapability
1MiMo-V2.6-Flash309B MoEMITcommunity59 tok/sagent-ready49/50
2Qwen3.8-Flash-Next125B MoEQwen Customestimated57 tok/sagent-ready47/50
3DeepSeek V4.1 Flash raise GPU limit (macOS 15)552B MoEMITcommunity20 tok/sagent-ready50/50
4GLM 5.2753B MoEMITcommunity15 tok/sagent-ready50/50
5MiniMax-M3428B MoEMiniMax Communitycommunity29 tok/sagent-ready48/50
6GLM-5.3 slow753B MoEGLM-5.3estimated11 tok/sagent-ready50/50
7Muse Spark Open816B MoEMuse Model Licensesourced18 tok/s—49/50
8MiniMax-M3.1-Flash428B MoEMiniMax Communitycommunity26 tok/sagent-ready48/50
9Inkling 2 Small594B MoEApache 2.0estimated25 tok/sagent-ready48/50
10KAT-Coder-V2.535B MoEApache 2.0estimated100 tok/sagent-ready39/50
11Nemotron 3.5 Lightning30B MoEOpenMDWestimated151 tok/sagent-ready36/50
12North Mini Code 1.030.5B MoEApache 2.0estimated133 tok/sagent-ready36/50
Provenance key
sourced
Published figure — vendor or press review
estimated
community
Linked individual run — blog, forum or submission

Showing the top 12 of 122 models that fit in 512 GB. See the full leaderboard or all benchmarks.

Estimated figures come from the bandwidth formula; against 233 measured runs its typical error is ±9% for dense models, ±15% for small MoEs (≤40B total) and ±34% for large MoEs (mostly below the measured figure). How the error is computed →

Context: “fits” means the weights at a short prompt. Coding agents (Claude Code, OpenCode, Codex CLI) hold 64K tokens or more, and every token adds KV cache: agent-ready marks a model whose weights plus a 64K-token 16-bit KV cache still fit in the 496 GB this Mac’s GPU can use with the limit raised (109 of the 111 models here with a KV figure). That is an estimate from each model’s config.json, not a measurement; a q8 KV cache halves the cache part, and — means the index has no KV figure for the model yet. The GPU memory limit below applies on top.

GPU memory limit: macOS lets the GPU use about 384 GB of this Mac’s 512 GB by default on macOS 15 and earlier, and about 445.4 GB on macOS 26 and later. Of the 122 models that fit, 5 need more than that at a short prompt and are tagged raise GPU limit. sudo sysctl iogpu.wired_limit_mb=507904 raises it to 496 GB, leaving 16 GB for macOS, until the Mac restarts (sudo sysctl iogpu.wired_limit_mb=0 restores the default). How the limit is read →

Quick start: run MiMo-V2.6-Flash on your Mac Studio M3 Ultra

The index has no verified Ollama tag for MiMo-V2.6-Flash, so start from its listing and copy the exact name there: Hugging Face repo → · model card →

GGUF builds run in LM Studio or llama.cpp; MLX builds run in LM Studio or mlx-lm.

Prefer a GUI? LM Studio gives you a one-click download and chat window. For step-by-step help see our Ollama install guide, or open the MiMo-V2.6-Flash on M3 Ultra page for its figures and setup.

Storage

The models a 512 GB Mac can hold are big files: DeepSeek V4.1 Flash, the largest that fits here, is about a 432 GB download at the build it needs, and a few models that size fill an internal drive. A fast external Thunderbolt 5 or USB4 SSD keeps a model library within reach: Thunderbolt 5 external SSDs on Amazon →

Amazon affiliate link: as an Amazon Associate, LLMCheck earns from qualifying purchases.

Same Mac with more memory, and the upgrade path

Upgrade or save: MiMo-V2.6-Flash on other Macs

Same pick, other Macs

Twice as fast: no Mac in the index runs MiMo-V2.6-Flash at twice this Mac’s community59 tok/s; the fastest figure is the Mac Studio M5 Ultra (256 GB) at community77 tok/s, $9,499 new (Apple list price, checked 2026-11-14): build-to-order from Apple (Amazon does not stock this configuration).

For less: the cheapest Mac that still runs MiMo-V2.6-Flash at 15+ tok/s is the Mac Studio M2 Ultra (192 GB) at estimated~54 tok/s, about $3,000 refurbished or used: search Amazon Renewed listings →.

Speeds are each Mac’s own figure from the index, with its provenance mark. Prices come from data/macs.json: Apple’s list price for a new Mac (dated), an approximate street price for a used or refurbished one. Amazon stock checked 2026-11-18. Amazon links are affiliate links: as an Amazon Associate, LLMCheck earns from qualifying purchases.

🛒 The current Mac with the same 512 GB

Apple no longer sells the Mac Studio M3 Ultra (512 GB) new. A current Mac Studio M5 Ultra (512 GB) runs the same models with 1228 GB/s of memory bandwidth instead of 819 GB/s, so they generate faster. The Mac Studio M5 Ultra (512 GB) is build-to-order from Apple; Amazon does not stock that configuration.

As an Amazon Associate, LLMCheck earns from qualifying purchases. Affiliate links cost you nothing extra and never influence our rankings.

Cite this page

LLMCheck. (2026). Best Local LLMs for the Mac Studio M3 Ultra (512 GB). llmcheck.net. https://llmcheck.net/best-llm/mac-studio-m3-ultra-512gb/ (updated ). CC BY 4.0.

How to cite LLMCheck →Data: CC BY 4.0Markdown version

FAQ: local LLMs on the Mac Studio M3 Ultra

What is the best local LLM for a Mac Studio M3 Ultra (512 GB)?

MiMo-V2.6-Flash (309B MoE, MIT) is the best all-round pick at 59 tok/s (community run) on the Mac Studio M3 Ultra. If you want maximum speed, LFM3-1.7B hits ~334 tok/s (estimated); for maximum capability, DeepSeek V4.1 Flash still fits in 512 GB.

How many models can a Mac Studio M3 Ultra with 512 GB run?

About 122 of the 130 models in the LLMCheck index fit in 512 GB of unified memory, from compact models up to Mistral Large 4 (1.05T MoE).

Can a Mac Studio M3 Ultra run a 70B model?

Yes. A dense 70B such as Llama 3.3 70B needs about 44 GB at 4-bit, which fits in 512 GB with room for context; the index puts it at ~16 tok/s (estimated) on this Mac.

Is 512 GB of RAM enough to run LLMs locally?

512 GB is enough for dense models up to about 405B at 4-bit (the largest in the index that fits is Llama 3.1 405B), and for mixture-of-experts models such as DeepSeek V4.1 Flash (20 tok/s (community run)). As a rule of thumb a 4-bit model needs about 0.57 GB per billion parameters plus ~2 GB, and macOS needs its own share of memory too. Because Apple Silicon uses unified memory, the same pool serves as system RAM and GPU memory.

Related