M5 Pro

Name: M5 Pro
Brand: Apple
Price: 59900 TWD

Apple M5 Pro: 24–64 GB of memory at 307 GB/s. Below: the open models it can run, estimated at 64 GB with an 8K context. 24–64 GB of unified memory at 307 GB/s in the MacBook Pro and Mac mini; 48 GB holds 32B models at 6-bit.

Last updated: 2026-10-14

Specifications

Memory	24–64 GB LPDDR5X[src]
Memory bandwidth	307 GB/s[src]
Launch	2026-03[src]
Launch price	US$2,199[src]
Architecture	Apple M5 Pro (Fusion Architecture: two third-generation 3 nm dies)[src]
Software	Metal / MLX

What it can run

Best-quality quantization that fits with a 8K context, and the estimated generation speed. Estimates, not benchmarks. See the methodology.

Memory configuration

Model	Params	Best quant	Memory	tok/s (est.)
Llama 3.3 70B Instruct	70B	Q4_K_M	44.5 GiB	4.7
Qwen3.5-35B-A3B	35B (3B)	Q8_0	37.0 GiB	63
Qwen3.6-35B-A3B	35B (3B)	Q8_0	37.0 GiB	63
Qwen3-32B	32.8B	Q8_0	36.6 GiB	5.7
DeepSeek-R1-Distill-Qwen-32B	32.76B	Q8_0	36.5 GiB	5.7
Gemma 4 31B IT	30.7B	Q8_0	33.5 GiB	6.1
Qwen3-30B-A3B (2507)	30.5B (3.3B)	Q8_0	32.9 GiB	57
Qwen3-Coder-30B-A3B-Instruct	30.5B (3.3B)	Q8_0	32.9 GiB	57
GLM-4.7-Flash (30B-A3B)	30B (3B)	Q8_0	32.1 GiB	63
Qwen3.5-27B	27B	Q8_0	29.1 GiB	7.0
Qwen3.6-27B	27B	Q8_0	29.1 GiB	7.0
Qwen3.8-27B	27B	Q8_0	29.1 GiB	7.0
Gemma 3 27B IT	27B	Q8_0	29.6 GiB	7.0
Gemma 4 26B A4B IT	25.2B (3.8B)	Q8_0	27.0 GiB	49
Mistral Small 3.2 24B Instruct (2506)	24B	Q8_0	26.7 GiB	7.8
Magistral Small 1.2 (2509)	24B	Q8_0	26.7 GiB	7.8
Devstral Small 2 24B Instruct (2512)	24B	Q8_0	26.7 GiB	7.8
gpt-oss-20b	21B (3.6B)	FP16	41.8 GiB	28
Phi-4 (14B)	14B	FP16	29.4 GiB	7.1
Gemma 3 12B IT	12B	FP16	24.8 GiB	8.3
Gemma 4 12B Unified IT	11.95B	FP16	24.2 GiB	8.3
Qwen3.5-9B	9B	FP16	18.4 GiB	11
Qwen3-8B	8.2B	FP16	17.7 GiB	12
DeepSeek-R1-0528-Qwen3-8B	8.19B	FP16	17.6 GiB	12
Llama 3.1 8B Instruct	8B	FP16	17.1 GiB	12
Gemma 4 E4B IT	8B	FP16	16.4 GiB	12
Llama 3.2 3B Instruct	3.21B	FP16	7.7 GiB	31
Kimi K2.6	1000B (32B)	Doesn't fit	–	–
DeepSeek-R1-0528 (671B)	671B (37B)	Doesn't fit	–	–
DeepSeek-V4-Flash	284B (13B)	Doesn't fit	–	–
Qwen3.8-Flash-Next	180B (6B)	Doesn't fit	–	–
Qwen3.5-122B-A10B	122B (10B)	Doesn't fit	–	–
Mistral Small 4 119B (2603)	119B (6.5B)	Doesn't fit	–	–
gpt-oss-120b	117B (5.1B)	Doesn't fit	–	–
Llama 4 Scout 17B-16E Instruct	109B (17B)	Doesn't fit	–	–
GLM-4.5-Air	106B (12B)	Doesn't fit	–	–
Qwen3-Coder-Next (80B-A3B)	80B (3B)	Doesn't fit	–	–

Prices

Latest price at each shop, as listed on the day we checked.

Region	Shop	Listing	Price	Per GB	Checked
Taiwan	Apple	Mac mini，M5 Pro 晶片，15 核心 CPU，16 核心 GPU，24GB 記憶體，512GB 儲存裝置Price shown in Apple Store configurator for this configuration	NT$59,900	NT$2,496/GB	2026-10-14
US	Apple	Mac mini, M5 Pro chip, 15-core CPU, 16-core GPU, 24GB memory, 512GB storage	US$1,699	US$71/GB	2026-10-14

Taiwan value: about 41.1 tok/s on Llama 3.1 8B Instruct (Q4_K_M, 8K), or 0.7 tok/s per NT$1,000.
US value: about 41.1 tok/s on Llama 3.1 8B Instruct (Q4_K_M, 8K), or 2.4 tok/s per US$100.

A price history chart appears once we have checked this device on two or more days.

All prices →

Frequently asked questions

Which LLMs can the M5 Pro run?

With 64 GB and an 8K context, the largest models that fit include: Llama 3.3 70B Instruct, Qwen3.5-35B-A3B, Qwen3.6-35B-A3B, Qwen3-32B, DeepSeek-R1-Distill-Qwen-32B. See the full table below.

How fast is the M5 Pro for local LLMs?

Generation speed is mostly set by memory bandwidth (307 GB/s). For Llama 3.1 8B Instruct at FP16, expect roughly 12 tokens/s; bigger models are slower.

Can the M5 Pro run 70B models?

Yes. Llama 3.3 70B Instruct fits on the M5 Pro (64 GB) at Q4_K_M with an 8K context, using about 44.5 GiB, at an estimated 4.7 tokens/s.