llmrun
AppleM5 ProMini PC

Best AI Models for Mac Mini M5 Pro (24 GB)

Memory:24 GB UnifiedBandwidth:307.0 GB/sGPU Cores:16 GPU coresCPU Cores:15 CPU coresNeural Engine:38.0 TOPS

24 GB unified − 3.5 GB OS overhead = 20.5 GB available for AI models

How does this device compare with your machine?

99 of 293 featured models run great on Mac Mini M5 Pro (24 GB) · best pick: Qwen3.8 27B at ~12.3 tok/s

largest that fits: 36.0B (Qwen3.6 35B A3B, IQ4_XS) · 70B needs CPU offload or more memory

What LLMs Can Mac Mini M5 Pro (24 GB) Run?

1068 models · 470 run great · 348 run well

Featured only

Score: model quality, 100 = average model. Fit: how well it runs on this hardware, S best to F won't fit. Ordered by quality, recency and Fit. How ranking works

Showing compatibility for Mac Mini M5 Pro (24 GB)

LLM models compatible with Mac Mini M5 Pro (24 GB) — ranked by quality, recency and Fit
ModelVRAMScoreFit
Gemma4 E4B MiniFantasy V1
8.0B·BF16·13.0 t/s·131K ctx
16.8 GBB49, Decent
ERNIE 4.5 0.3B Paddle
361M·BF16·206.6 t/s·131K ctx
1.2 GBS100, Runs great
Huihui Qwen3 4B Abliterated v2
4.0B·Q4_K_M·74.1 t/s·41K ctx
4.2 GBS86, Runs great
Qwen2.5 3B Instruct
3.1B·Q4_K_M·96.4 t/s·33K ctx
2.8 GBS90, Runs great
Huihui Qwen3 8B Abliterated v2
8.2B·Q4_K_M·38.9 t/s·41K ctx
7.6 GBA75, Runs well
Tiny Random Llama 3
4M·Q4_K_M·716.3 t/s·131K ctx
0.3 GBS100, Runs great
OTel LLM 27B IT
27B·Q4_K_M·12.0 t/s·131K ctx
19.3 GBC30, Usable
Phi 4 Mini Flash Reasoning
3.9B·Q4_K_M·72.8 t/s·262K ctx
5.3 GBS86, Runs great
Llama 3.2 1B
1.2B·Q4_K_M·193.6 t/s·131K ctx
1.6 GBS99, Runs great
Gemma 4 E4B Luchador
8.0B·BF16·13.0 t/s·131K ctx
16.8 GBB49, Decent
EuroLLM 1.7B Instruct
1.7B·Q4_K_M·143.3 t/s·4K ctx
1.7 GBS96, Runs great
T5gemma L L Ul2 IT
1.2B·BF16·78.7 t/s
4.3 GBS87, Runs great
H2ovl Mississippi 2B
2.2B·BF16·45.3 t/s
6.5 GBA77, Runs well
Gemma 3 1B IT
1000M·Q4_K_M·325.6 t/s·33K ctx
2.1 GBS100, Runs great
Tiny Dummy Qwen2
1M·Q4_K_M·716.3 t/s·33K ctx
0.5 GBS100, Runs great
Schematron 8B
8B·Q4_K_M·40.0 t/s·131K ctx
7.3 GBA75, Runs well

Mac Mini M5 Pro (24 GB) has 20.5 GB of memory for models. In the ranking below, Qwen3.8 27B leads at Q4_K_M: it needs 17.9 GB (87% of that memory) and generates about 12.3 tokens per second. 99 of the 293 featured models earn a Fit S here, meaning they load with room for context and run fast. Decode speed is set mostly by memory bandwidth (307 GB/s), so smaller and mixture-of-experts models are noticeably quicker than dense models of the same quality. 70B-class models need about 46 GB at Q4_K_M, more than the 20.5 GB available, so they require offloading layers to system RAM or a larger-memory setup.

Runs Well

  • Gemma 4 26B A4B IT (25.8B at Q4_K_M, ~53.2 tok/s)
  • GPT OSS 20B (20.9B at Q4_K_M, ~59.3 tok/s)
  • Qwen3.5 9B (9.7B at Q4_K_M, ~33.8 tok/s)
  • MiniCPM5 2B (2.5B at Q4_K_M, ~113 tok/s)
  • Ling 3.0 Tiny (7.9B at Q4_K_M, ~109 tok/s)

Challenging

  • Qwen3.8 27B (27.8B at Q4_K_M, ~12.3 tok/s) fits but uses 87% of memory, so keep context short
  • Qwen3.6 35B A3B (36.0B at IQ4_XS, ~49.7 tok/s) fits but uses 97% of memory, so keep context short
  • NVIDIA Nemotron 3.5 Lightning 30B A3B BF16 (31.6B at Q4_K_M, ~58.4 tok/s) fits but uses 94% of memory, so keep context short
  • 70B-class models need ~46 GB at Q4_K_M, so they need CPU offload or more memory
  • 120B-class models need ~77 GB at Q4_K_M, so they need CPU offload or more memory

Mac Mini M5 Pro (24 GB) Specifications

Brand
Apple
Chip
M5 Pro
Type
Mini PC
Unified Memory
24 GB
Memory Bandwidth
307.0 GB/s
GPU Cores
16
CPU Cores
15
Neural Engine
38.0 TOPS
Release Date
2026-09-22

Get Started

Ollama (Recommended)

Runs Qwen3.8 27B on Q4_K_M; this is our top choice here.

$curl -fsSL https://ollama.com/install.sh | sh
$ollama run hf.co/unsloth/Qwen3.8-27B-GGUF:Q4_K_M

LM Studio

LM Studio

Download LM Studio, search for "Qwen3.8 27B", and run it with one click.

Prompt Processing

Estimated for Qwen3.8 27B, the compute-bound phase that reads your prompt before the first reply token appears.

93tok/s prefill

Short chat

5.5 s

512 tok prompt

Long chat

44.0 s

4,096 tok prompt

Document / codebase

352 s

32,768 tok prompt

Prefill is compute-bound and a different number from the decode tok/s shown elsewhere on this page — how prompt processing works →

Performance figures are estimates calibrated as of 2026-09-21 — see calibration basis →

What to do next with Mac Mini M5 Pro (24 GB)

Which model to run first, a faster option with the same memory, and what more memory unlocks. Computed from the rankings and specs above.

Frequently Asked Questions

Can Mac Mini M5 Pro (24 GB) run Qwen3.8 27B?

Yes. The Mac Mini M5 Pro (24 GB) with 24 GB unified memory can run Qwen3.8 27B (Q4_K_M, 17.9 GB, 87% of memory). It is also the table's best pick, at ~12.3 tok/s. 99 of the 293 featured models get a Fit of S (runs great) here. The largest featured model that fits at Q4 or better is Qwen3.6 35B A3B (36.0B at IQ4_XS). 70B-class models need around 46 GB at Q4_K_M, meaning they need CPU offloading or extra memory beyond what's available in 20.5 GB. Check the compatibility table above for the full list with Fit, VRAM usage and estimated speed. Models are checked against the 20.5 GB available to AI (that is, the 24 GB of RAM minus roughly 3.5 GB for macOS), while Apple Silicon's unified memory allows the GPU to draw from that same pool without copying data.

How much memory is available for AI on Mac Mini M5 Pro (24 GB)?

The Mac Mini M5 Pro (24 GB) has 24 GB unified memory. After macOS reserves ~3.5 GB for the operating system, approximately 20.5 GB is available for AI models. Model rankings here use that 20.5 GB limit instead of the total 24 GB. Unlike discrete GPUs where VRAM is separate from system RAM, a unified-memory device shares one memory pool between the CPU and GPU — this means no data copying overhead, but you share memory with macOS and open apps.

Is Mac Mini M5 Pro (24 GB) good for AI?

With 24 GB unified memory and 307.0 GB/s bandwidth, the Mac Mini M5 Pro (24 GB) is solid for running local AI models. Models that run well include Gemma 4 26B A4B IT (25.8B at Q4_K_M, ~53.2 tok/s); GPT OSS 20B (20.9B at Q4_K_M, ~59.3 tok/s); Qwen3.5 9B (9.7B at Q4_K_M, ~33.8 tok/s). Mid-size models run at good quality, and Metal acceleration keeps interactive chat smooth.

What's the best model for Mac Mini M5 Pro (24 GB)?

For the Mac Mini M5 Pro (24 GB), the best pick right now is Qwen3.8 27B at ~12.3 tok/s. Also running well: Gemma 4 26B A4B IT (25.8B at Q4_K_M, ~53.2 tok/s); GPT OSS 20B (20.9B at Q4_K_M, ~59.3 tok/s); Qwen3.5 9B (9.7B at Q4_K_M, ~33.8 tok/s). The best choice depends on your use case: coding assistants benefit from code-tuned models, general chat from instruction-tuned ones. The list is ranked by quality, recency and Fit.

How fast is Mac Mini M5 Pro (24 GB) for AI inference?

With 307.0 GB/s memory bandwidth, the Mac Mini M5 Pro (24 GB) achieves approximately 48 tok/s on a 7B model at Q4_K_M — that's very fast, well above conversational speed. A 14B model runs at ~24 tok/s. In the ranking, the top pick Qwen3.8 27B runs at ~12.3 tok/s. Apple Silicon achieves high efficiency (~70%) thanks to unified memory — there's no PCIe bottleneck between CPU and GPU.

= (307 GB/s ÷ model GB) × efficiency

Apple Silicon achieves ~70% bandwidth efficiency thanks to and Metal acceleration.

Estimated speed on Mac Mini M5 Pro (24 GB)

~12 tok/s
~53 tok/s
~50 tok/s
~59 tok/s

Real-world results typically within ±20%.

Learn more about tok/s estimation →

Can I run AI offline on Mac Mini M5 Pro (24 GB)?

Yes — once you download a model, it runs entirely on the Mac Mini M5 Pro (24 GB) without internet. Applications like Ollama and LM Studio make it straightforward to download, manage, and run models locally. All your conversations stay private on your device with zero data sent to external servers. This is one of the key advantages of local AI: complete privacy, no API costs, and no rate limits.