What can a MacBook Pro M2 Max 96GB run?

96GB · 400 GB/s · 90W

The 'good enough Mac' — 96 GB of unified memory that still chews through 32B Q8 and 70B Q3.

Runs great27

Comfortable headroom — daily-driver territory.

Model	Quant	Est. speed	Max context
deepseek moe 16b chat	FP16	Infinity tok/s	—
qwen2 7b instruct	FP16	Infinity tok/s	—
Qwen3-0.6B	FP16	200 tok/s	33k
Gemma-3n-E4B	FP16	88 tok/s	131k
SmolLM2-1.7B	FP16	73 tok/s	8k
Qwen3-30B-A3B	Q8_0	71 tok/s	33k
Qwen3-Coder-30B-A3B	Q8_0	71 tok/s	231k
Llama-3.2-3B	FP16	39 tok/s	131k
GPT-OSS-20B	FP16	34 tok/s	131k
Phi-4-mini	FP16	32 tok/s	131k
Gemma-3-4B	FP16	32 tok/s	131k
Qwen3-4B	FP16	32 tok/s	33k
Gemma-3-12B	Q8_0	18 tok/s	131k
Qwen3-14B	Q6_K	18 tok/s	33k
Mistral-7B-Instruct	FP16	17 tok/s	33k
Mistral-Small-24B	Q4_K_M	17 tok/s	131k
DeepSeek-R1-Distill-Qwen-7B	FP16	16 tok/s	131k
Qwen2.5-Coder-7B	FP16	16 tok/s	131k
Phi-4-14B	Q8_0	16 tok/s	16k
Codestral-22B	Q5_K_M	16 tok/s	33k
DeepSeek-R1-Distill-Qwen-14B	Q8_0	16 tok/s	131k
Llama-3.1-8B	FP16	15 tok/s	131k
Qwen3-8B	FP16	15 tok/s	33k
Gemma-3-27B	Q4_K_M	15 tok/s	131k
Qwen3.8	Q4_K_M	15 tok/s	262k
Qwen3.8 27B	Q4_K_M	15 tok/s	262k
qwen35 27B	Q4_K_M	15 tok/s	341k

Tight squeeze8

It fits, but close the browser tabs and mind the context.

Model	Quant	Est. speed	Max context
Qwen3.8 Flash Next	IQ2_M	245 tok/s	42k
GPT-OSS-120B	MLX 4-bit	86 tok/s	14k
Llama-3.3-70B	Q6_K	4 tok/s	56k
Qwen2.5-72B	Q6_K	4 tok/s	46k
glm-4.6-air 106B	Q4_K_M	4 tok/s	22k
DeepSeek-R1-Distill-Qwen-32B	FP16	4 tok/s	19k
GLM-4-32B	FP16	4 tok/s	19k
Qwen3-32B	FP16	4 tok/s	19k

Measured on this card0

Everything here was benchmarked on this exact card by someone who owns one — the tables above are our arithmetic; these numbers came off real machines.

Nobody has measured anything on this card yet. Paste a benchmark — it takes one keystroke and no form.

Try your own machine →← All hardware