AI Models for MacBook Pro M5 Max — What runs on 128GB
Specs checked against manufacturer and public documentation on . Results below are estimates, not measurements.
What runs on the MacBook Pro M5 Max
All 66 models at the best tracked quant for this 128GB configuration. Select any row for the full report and step-by-step setup guide.
| Model | Params | Quant | Needs | Speed | Verdict |
|---|---|---|---|---|---|
| Hunyuan 3 (Hy3) | 298.8B | IQ1_Ssmaller version | 89GB | ~4.8–6.7 tokens/s | ✓ Runs great |
| Laguna S 2.1 | 118B | Q4_K_M | 80.7GB | ~66.4–93 tokens/s | ✓ Runs great |
| Llama 3.3 70B | 70B | Q4_K_M | 50.3GB | ~7.2–10.1 tokens/s | ✓ Runs great |
| Qwen 3.5 35B-A3B | 35B | Q4_K_M | 26.4GB | ~162.8–227.9 tokens/s | ✓ Runs great |
| Qwen 3.6 35B-A3B | 35B | Q4_K_M | 24.1GB | ~162.1–226.9 tokens/s | ✓ Runs great |
| Ornith 1.0 35B-A3B | 35B | Q4_K_M | 25.5GB | ~168.9–236.5 tokens/s | ✓ Runs great |
| KAT-Coder V2.5 Dev | 35B | Q4_K_M | 25.7GB | ~167.4–234.3 tokens/s | ✓ Runs great |
| Laguna XS 2.1 | 33B | Q4_K_M | 24.4GB | ~166.4–232.9 tokens/s | ✓ Runs great |
| Qwen3 32B | 32.8B | Q4_K_M | 23.9GB | ~15.5–21.7 tokens/s | ✓ Runs great |
| Gemma 4 31B | 31B | Q4_K_M | 22.2GB | ~16.8–23.5 tokens/s | ✓ Runs great |
| Qwen3 30B A3B | 30.5B | Q4_K_M | 22.5GB | ~152.5–213.6 tokens/s | ✓ Runs great |
| Nemotron 3 Nano 30B-A3B | 30B | Q4_K_M | 28.7GB | ~124.8–174.7 tokens/s | ✓ Runs great |
| Salience 1.5 Flash | 30B | Q4_K_M | 22.5GB | ~149.2–208.9 tokens/s | ✓ Runs great |
| Qwen 3.5 27B | 27B | Q4_K_M | 20.2GB | ~18.4–25.7 tokens/s | ✓ Runs great |
| Qwen 3.6 27B | 27B | Q4_K_M | 18.7GB | ~18.3–25.6 tokens/s | ✓ Runs great |
| Bonsai 27B (1-bit) | 27B | Q1_0 | 5GB | ~80.8–113.1 tokens/s | ✓ Runs great |
| Ternary Bonsai 27B | 27B | Q2_0 | 8.6GB | ~42.6–59.7 tokens/s | ✓ Runs great |
| Fara 1.5 27B | 27B | Q4_K_M | 21.1GB | ~17.5–24.6 tokens/s | ✓ Runs great |
| Gemma 4 26B-A4B | 26B | Q4_0 | 17.7GB | ~138.6–194 tokens/s | ✓ Runs great |
| GPT-OSS 20B | 21B | MXFP4 | 15GB | ~148–207.2 tokens/s | ✓ Runs great |
| Qwen3 14B | 14.8B | Q4_K_M | 11.3GB | ~34.1–47.8 tokens/s | ✓ Runs great |
| Phi-4 14B | 14.7B | Q4_K_M | 11.4GB | ~33.7–47.2 tokens/s | ✓ Runs great |
| Ministral 3 14B | 14B | Q4_K_M | 10.4GB | ~37.4–52.4 tokens/s | ✓ Runs great |
| Gemma 3 12B | 12.2B | Q4_K_M | 9.3GB | ~42.1–58.9 tokens/s | ✓ Runs great |
| Gemma 4 12B | 12B | Q4_0 | 9GB | ~43.9–61.4 tokens/s | ✓ Runs great |
| Grug 12B | 12B | Q4_K_M | 9.7GB | ~39.9–55.8 tokens/s | ✓ Runs great |
| grug 9B (ProCreations) | 9.4B | Q4_K_M | 7.7GB | ~52–72.8 tokens/s | ✓ Runs great |
| Fara 1.5 9B | 9.4B | Q4_K_M | 7.7GB | ~52–72.8 tokens/s | ✓ Runs great |
| Qwen 3.5 9B | 9B | Q4_K_M | 7.4GB | ~53.9–75.4 tokens/s | ✓ Runs great |
| Ornith 1.0 9B | 9B | Q4_K_M | 7.3GB | ~54.8–76.8 tokens/s | ✓ Runs great |
| Qwythos 9B v2 | 9B | Q4_K_M | 7.4GB | ~53.9–75.4 tokens/s | ✓ Runs great |
| Tini Cybersec 8B-A1B | 8.5B | Q4_K_M | 6.9GB | ~501.8–702.6 tokens/s | ✓ Runs great |
| Qwen3 8B | 8.2B | Q4_K_M | 6.6GB | ~61.4–86 tokens/s | ✓ Runs great |
| Llama 3.1 8B | 8B | Q4_K_M | 6.5GB | ~62.7–87.7 tokens/s | ✓ Runs great |
| Ministral 8B | 8B | Q4_K_M | 6.5GB | ~62.7–87.7 tokens/s | ✓ Runs great |
| Ternary Bonsai 8B | 8B | Q2_0 | 3.7GB | ~139.5–195.4 tokens/s | ✓ Runs great |
| Ministral 3 8B | 8B | Q4_K_M | 6.8GB | ~59–82.7 tokens/s | ✓ Runs great |
| LFM2.5 8B-A1B | 8B | Q4_K_M | 6.8GB | ~472.3–661.2 tokens/s | ✓ Runs great |
| DeepSeek R1 Distill 7B | 7.6B | Q4_K_M | 6.3GB | ~65.3–91.4 tokens/s | ✓ Runs great |
| Mistral 7B v0.3 | 7.2B | Q4_K_M | 5.9GB | ~69.8–97.7 tokens/s | ✓ Runs great |
| AREX Turbo 4B | 4.5B | Q4_K_M | 4.2GB | ~105.9–148.2 tokens/s | ✓ Runs great |
| Fara 1.5 4B | 4.5B | Q4_K_M | 4.2GB | ~105.9–148.2 tokens/s | ✓ Runs great |
| Gemma 3 4B | 4.3B | Q4_K_M | 3.7GB | ~122.8–171.9 tokens/s | ✓ Runs great |
| Qwen3 4B | 4B | Q4_K_M | 3.7GB | ~122.8–171.9 tokens/s | ✓ Runs great |
| Qwen 3.5 4B | 4B | Q4_K_M | 3.9GB | ~113.7–159.2 tokens/s | ✓ Runs great |
| Gemma 4 E4B | 4B | Q4_K_M | 6.3GB | ~61.4–86 tokens/s | ✓ Runs great |
| Ternary Bonsai 4B | 4B | Q2_0 | 2.2GB | ~279.1–390.7 tokens/s | ✓ Runs great |
| Nemotron 3 Nano 4B | 4B | Q4_K_M | 4GB | ~109.6–153.5 tokens/s | ✓ Runs great |
| Agents-A1 4B | 4B | Q4_K_M | 3.9GB | ~113.7–159.2 tokens/s | ✓ Runs great |
| Phi-4 Mini 3.8B | 3.8B | Q4_K_M | 3.7GB | ~122.8–171.9 tokens/s | ✓ Runs great |
| Llama 3.2 3B | 3.2B | Q4_K_M | 3.1GB | ~153.5–214.9 tokens/s | ✓ Runs great |
| SmolLM3 3B | 3.1B | Q4_K_M | 3GB | ~161.6–226.2 tokens/s | ✓ Runs great |
| Ministral 3 3B | 3B | Q4_K_M | 3.2GB | ~146.2–204.7 tokens/s | ✓ Runs great |
| G9v3 3B | 3B | Q4_K_M | 3GB | ~161.6–226.2 tokens/s | ✓ Runs great |
| Qwen 3.5 2B | 2B | Q4_K_M | 2.3GB | ~236.2–330.6 tokens/s | ✓ Runs great |
| Gemma 4 E2B | 2B | Q4_K_M | 4.2GB | ~99–138.6 tokens/s | ✓ Runs great |
| DeepSeek R1 Distill 1.5B | 1.8B | Q4_K_M | 2.1GB | ~279.1–390.7 tokens/s | ✓ Runs great |
| Qwen3 1.7B | 1.7B | Q8_0 | 2.8GB | ~170.6–238.8 tokens/s | ✓ Runs great |
| SmolLM2 1.7B | 1.7B | Q4_K_M | 2.1GB | ~279.1–390.7 tokens/s | ✓ Runs great |
| Ternary Bonsai 1.7B | 1.7B | Q2_0 | 1.4GB | ~614–859.6 tokens/s | ✓ Runs great |
| Llama 3.2 1B | 1.2B | Q4_K_M | 1.7GB | ~383.8–537.2 tokens/s | ✓ Runs great |
| Gemma 3 1B | 1B | Q4_K_M | 1.7GB | ~383.8–537.2 tokens/s | ✓ Runs great |
| OvisOCR2 0.8B | 0.8B | Q4_K_M | 1.5GB | ~511.7–716.3 tokens/s | ✓ Runs great |
| Qwen3 0.6B | 0.6B | Q8_0 | 1.5GB | ~511.7–716.3 tokens/s | ✓ Runs great |
| Inkling | 952.4B | Q8_0 | 967.1GB | — | ✕ Won't fit |
| Ornith 1.0 397B | 397B | Q4_K_M | 285.8GB | — | ✕ Won't fit |
~ = bandwidth-based estimate · open a row to see the recommended app and exact steps
Other MacBook Pro M5 Max memory options
FAQ
What is the biggest local AI model the MacBook Pro M5 Max · 128GB can run?
Hunyuan 3 (Hy3) is the largest model in our 66-model comparison that fits at IQ1_S. It needs about 89GB inside our 120GB working-memory budget, with an estimated 4.8–6.7 tokens/s decode range.
How much of the 128GB unified memory is available to a local LLM?
We use a conservative 120GB working budget, leaving room for macOS, the inference app, and normal background activity. Memory pressure, context length, and other open apps can change the real limit.
Can the MacBook Pro M5 Max · 128GB run Llama 3.1 8B?
Yes. At Q4_K_M, our estimate uses 6.5GB and lands around 62.7–87.7 tokens/s.
Can the MacBook Pro M5 Max · 128GB run Qwen 3.6 27B?
Yes at Q4_K_M: about 18.7GB of working memory and an estimated 18.3–25.6 tokens/s.
Which app should I use for local AI on the MacBook Pro M5 Max?
For an exact GGUF repository and quant from this site, start with LM Studio's graphical Discover, download, load, and chat flow. Jan is the open-source GGUF alternative; Ollama is strongest when the exact model already has a trustworthy catalog package or you want coding/API integrations; Msty is useful for mixed GGUF, MLX, and document workflows. For the curated Bonsai 27B build, try Locally AI and first confirm that its catalog shows the model on this Mac.
Can I upgrade the MacBook Pro M5 Max to more unified memory later?
No. Apple-silicon unified memory is integrated into the chip package and must be chosen at purchase. Storage upgrades or external SSDs do not increase the memory an LLM can use.
Are these MacBook Pro M5 Max local AI speeds measured?
No. Every speed on this page is a formula range based on Apple’s published memory bandwidth, model size, active parameters, and a cooling-aware efficiency range. App, backend, context length, and thermals can change real results.