M6 (24GB/32GB configurations, 170 GB/s)

Name: M6 (24GB/32GB configurations, 170 GB/s)
Brand: Apple
Price: 36900 TWD

Apple M6 (24GB/32GB configurations, 170 GB/s): 24–32 GB of memory at 170 GB/s. Below: the open models it can run, estimated at 32 GB with an 8K context. The same M6 with 24 or 32 GB runs at 170 GB/s instead of 153 GB/s; 32 GB fits MoE models such as Qwen3.6-35B-A3B at 4-bit.

Last updated: 2026-10-14

Specifications

Memory	24–32 GB[src]
Memory bandwidth	170 GB/s[src]
Launch	2026-09[src]
Architecture	Apple M6 (2 nm)[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.)
Qwen3.5-35B-A3B	35B (3B)	Q4_K_M	21.4 GiB	61
Qwen3.6-35B-A3B	35B (3B)	Q4_K_M	21.4 GiB	61
Qwen3-32B	32.8B	Q4_K_M	21.9 GiB	5.6
DeepSeek-R1-Distill-Qwen-32B	32.76B	Q4_K_M	21.9 GiB	5.6
Gemma 4 31B IT	30.7B	Q5_K_M	22.9 GiB	5.1
Qwen3-30B-A3B (2507)	30.5B (3.3B)	Q5_K_M	22.5 GiB	47
Qwen3-Coder-30B-A3B-Instruct	30.5B (3.3B)	Q5_K_M	22.5 GiB	47
GLM-4.7-Flash (30B-A3B)	30B (3B)	Q5_K_M	21.8 GiB	52
Qwen3.5-27B	27B	Q6_K	22.7 GiB	5.0
Qwen3.6-27B	27B	Q6_K	22.7 GiB	5.0
Qwen3.8-27B	27B	Q6_K	22.7 GiB	5.0
Gemma 3 27B IT	27B	Q6_K	23.3 GiB	5.0
Gemma 4 26B A4B IT	25.2B (3.8B)	Q6_K	21.1 GiB	35
Mistral Small 3.2 24B Instruct (2506)	24B	Q6_K	21.1 GiB	5.6
Magistral Small 1.2 (2509)	24B	Q6_K	21.1 GiB	5.6
Devstral Small 2 24B Instruct (2512)	24B	Q6_K	21.1 GiB	5.6
gpt-oss-20b	21B (3.6B)	Q8_0	22.5 GiB	29
Phi-4 (14B)	14B	Q8_0	16.6 GiB	7.4
Gemma 3 12B IT	12B	Q8_0	13.8 GiB	8.7
Gemma 4 12B Unified IT	11.95B	Q8_0	13.3 GiB	8.7
Qwen3.5-9B	9B	FP16	18.4 GiB	6.1
Qwen3-8B	8.2B	FP16	17.7 GiB	6.7
DeepSeek-R1-0528-Qwen3-8B	8.19B	FP16	17.6 GiB	6.7
Llama 3.1 8B Instruct	8B	FP16	17.1 GiB	6.9
Gemma 4 E4B IT	8B	FP16	16.4 GiB	6.9
Llama 3.2 3B Instruct	3.21B	FP16	7.7 GiB	17
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	–	–
Llama 3.3 70B Instruct	70B	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，M6 晶片，12 核心 CPU，12 核心 GPU，24GB 記憶體，256GB 儲存裝置Price shown in Apple Store configurator for this configuration	NT$36,900	NT$1,538/GB	2026-10-14
US	Apple	Mac mini, M6 chip, 12-core CPU, 12-core GPU, 24GB memory, 256GB storagePrice shown in Apple Store configurator for this configuration	US$1,099	US$46/GB	2026-10-14

Taiwan value: about 22.8 tok/s on Llama 3.1 8B Instruct (Q4_K_M, 8K), or 0.6 tok/s per NT$1,000.
US value: about 22.8 tok/s on Llama 3.1 8B Instruct (Q4_K_M, 8K), or 2.1 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 M6 (24GB/32GB configurations, 170 GB/s) run?

With 32 GB and an 8K context, the largest models that fit include: Qwen3.5-35B-A3B, Qwen3.6-35B-A3B, Qwen3-32B, DeepSeek-R1-Distill-Qwen-32B, Gemma 4 31B IT. See the full table below.

How fast is the M6 (24GB/32GB configurations, 170 GB/s) for local LLMs?

Generation speed is mostly set by memory bandwidth (170 GB/s). For Llama 3.1 8B Instruct at FP16, expect roughly 6.9 tokens/s; bigger models are slower.

Can the M6 (24GB/32GB configurations, 170 GB/s) run 70B models?

Llama 3.3 70B Instruct does not fit on the M6 (24GB/32GB configurations, 170 GB/s) (32 GB), even at Q4_K_M with an 8K context.