RunsLocalGPU checker Local AI toolsv2.2 · data Oct 17, 2026

What AI models can the M5 Max 128GB run?

With 128 GB unified memory, the M5 Max 128GB runs 21 of 25 popular open models fully in fast memory at Q4_K_M or better (8K context). The largest that fit are Qwen3.5 122B-A10B (Q5_K_M), gpt-oss-120b (Q5_K_M), Llama 4 Scout (Q6_K). Fastest: Gemma 4 E2B at over 150 tokens/sec and Mellum2.1 12B-A2.5B Thinking at about 139 tokens/sec.

Check with your own RAM and context length

Model	Result	Best quant	Memory needed	Est. speed
Gemma 4 E2B small	Runs on GPU	Q8_0	6.1 GB	150+ tok/s
Mellum2.1 12B-A2.5B Thinking MoE · coding	Runs on GPU	Q8_0	13.3 GB	139 tok/s
Nemotron 3.5 Lightning 30B-A3B MoE · agentic	Runs on GPU	Q8_0	32.7 GB	116 tok/s
Qwen3.6 35B-A3B MoE · coding	Runs on GPU	Q8_0	38.2 GB	116 tok/s
gpt-oss-120b MoE · reasoning	Runs on GPU	Q5_K_M	84.3 GB	102 tok/s
gpt-oss-20b MoE · reasoning	Runs on GPU	Q8_0	23.2 GB	96 tok/s
Ministral 3 3B small	Runs on GPU	Q8_0	4.8 GB	91 tok/s
Gemma 4 26B A4B MoE · vision	Runs on GPU	Q8_0	27.9 GB	91 tok/s
Gemma 4 E4B small · audio	Runs on GPU	Q8_0	9.3 GB	77 tok/s
Qwen3.5 122B-A10B MoE	Runs on GPU	Q5_K_M	88.1 GB	52 tok/s
Qwen3.5 9B general	Runs on GPU	Q8_0	11.3 GB	36 tok/s
Gemma 4 12B general · audio	Runs on GPU	Q8_0	13.8 GB	29 tok/s
Llama 4 Scout MoE · vision	Runs on GPU	Q6_K	91.4 GB	26 tok/s
Phi-4 14B reasoning	Runs on GPU	Q8_0	17.7 GB	24 tok/s
Mistral Small 3.1 24B general	Runs on GPU	Q8_0	27.3 GB	14 tok/s
Qwen3.6 27B coding	Runs on GPU	Q8_0	29.8 GB	13 tok/s
Qwen3.8 27B coding	Runs on GPU	Q8_0	30.6 GB	12 tok/s
Granite 4.2 30B general	Runs on GPU	Q8_0	31.9 GB	12 tok/s
Gemma 4 31B general · vision	Runs on GPU	Q8_0	33.8 GB	11 tok/s
Hermes 4.3 36B general	Runs on GPU	Q8_0	40.5 GB	9.6 tok/s
Llama 3.3 70B general	Runs on GPU	Q8_0	78.1 GB	4.9 tok/s
Qwen3.8-Flash-Next MoE	Lower quality fit	Q3_K_M	87.1 GB	126 tok/s
GLM-5.3-Flash MoE · coding	Too big	Q4_K_M	196.0 GB	—
DeepSeek V4.1-Flash MoE · coding	Too big	Q4_K_M	335.8 GB	—
Muse Spark Open 816B-A42B MoE · vision	Too big	Q4_K_M	497.3 GB	—

Specs used: 128 GB unified memory, 614 GB/s memory bandwidth. Speeds are estimates for generating text with llama.cpp-style runtimes; real numbers usually land within about 20%.

Apps that support the M5 Max 128GB

Ollama LM Studio Jan GPT4All AnythingLLM llama.cpp MLX LM KoboldCpp Text Generation WebUI LocalAI