Hardware

Apple M2 Max

Apple M2 Max has 32 / 64 / 96 GB Unified and 400 GB/s. See which open models fit, at which quant, and roughly how fast. Figures below use the default memory (32 GB). Updated 2026-10-24.

Memory32 / 64 / 96 GB Unified

Bandwidth400 GB/s

Kindmac

Year2023

400 GB/s. 32, 64, or 96GB. Source.

RSS for new models on this device

Models that fit

Model	Verdict	Quant	Context	tok/s
Swift Qwen3.8 27B 27.3B	Runs	Q5_K_M	8K	11–14
Qwen3.6 27B 26.9B	Tight fit	Q6_K	8K	9.5–13
Qwen3.5 27B 26.9B	Tight fit	Q6_K	8K	9.6–13
Kwaipilot KAT Coder V2.5 Dev 34.7B	Tight fit	Q5_K_S	8K	48–66
Devstral Small 2 24B Instruct 2512 23.6B	Tight fit	UD-Q6_K_XL	8K	9.9–14
Mistral Small 3.2 24B Instruct 2506 23.6B	Tight fit	UD-Q6_K_XL	8K	9.9–14
Magistral Small 2509 23.6B	Tight fit	UD-Q6_K_XL	8K	9.9–14
Qwen3.5 35B A3B 34.7B	Tight fit	UD-Q4_K_XL	8K	105–143
Qwen3.6 35B A3B 34.7B	Tight fit	UD-Q4_K_XL	8K	105–143
Tiel Coder 35B A3B 34.7B	Tight fit	UD-Q4_K_XL	8K	105–143
Nemotron 3 Nano 30B A3B 31.6B	Tight fit	Q5_K_S	8K	81–111
GLM 4.7 Flash 29.9B	Tight fit	UD-Q5_K_XL	8K	44–59
Cyber Tiel Coder 35B A3B GGUF MTP 35.5B	Tight fit	Q4_K_M (est.)	8K	111–151
Qwen3.8 35B A3B Distill 35.5B	Tight fit	Q4_K_M	8K	110–150
Qwen3 Coder 30B A3B Instruct 30.5B	Tight fit	UD-Q5_K_XL	8K	75–102
Qwen3 30B A3B Instruct 2507 30.5B	Tight fit	UD-Q5_K_XL	8K	75–102
Qwen3 VL 30B A3B Instruct 30.5B	Tight fit	UD-Q5_K_XL	8K	75–102
Qwen3 30B A3B Thinking 2507 30.5B	Tight fit	UD-Q5_K_XL	8K	75–102
DeepSeek R1 Distill Llama 70B 70.6B	Tight fit	UD-IQ2_XXS	8K	9.9–14
Llama 3.3 70B Instruct 70.6B	Tight fit	UD-IQ2_XXS	8K	9.9–14
Hermes 4 70B 70.6B	Tight fit	UD-IQ2_XXS	8K	9.9–14
NVIDIA Nemotron 3 Nano Omni 30B A3B Reasoning 31.6B	Tight fit	UD-Q4_K_XL	8K	81–111
NVIDIA Nemotron 3.5 Lightning 30B A3B 32.9B	Tight fit	UD-Q4_K_S	4K	90–123
bottlecapai ThinkingCap gemma 4 26B A4B it 25.2B	Runs	Q5_K_M	8K	42–58
gemma 4 26B A4B it 25.2B	Tight fit	UD-Q6_K_XL	8K	39–53
Qwen3 Coder Next 79.7B	Tight fit	UD-IQ2_XXS	8K	46–63
LFM2.5 24B A2B 24.0B	Tight fit	UD-Q6_K_XL	8K	88–120
Qwen3 Next 80B A3B Thinking 79.7B	Tight fit	UD-IQ1_M	4K	167–228
Qwen3 Next 80B A3B Instruct 79.7B	Tight fit	UD-IQ1_S	8K	132–180
GLM 4.7 Flash REAP 23B A3B 23.0B	Tight fit	Q8_0	4K	64–88
gpt oss 20b 20.9B	Runs great	UD-Q8_K_XL	8K	55–74
Qwen2.5 14B Instruct 14.8B	Runs	Q8_0	8K	13–17
NousResearch Hermes 4 14B 14.8B	Runs	Q8_0	8K	13–18
DeepSeek R1 Distill Qwen 14B 14.8B	Runs	Q8_0	8K	13–17
Qwen2.5 Coder 14B Instruct 128K 14.8B	Runs	Q8_0	8K	13–17
Qwen3 14B 14.8B	Runs	UD-Q8_K_XL	8K	11–15
phi 4 14.7B	Runs	Q8_0	8K	13–17
Ministral 3 14B Instruct 2512 13.5B	Runs	UD-Q8_K_XL	8K	12–16
Ministral 3 14B Reasoning 2512 13.5B	Runs	UD-Q8_K_XL	8K	12–16
Qwen3.6 14B A3B FableVibes 13.8B	Runs great	Q8_0	8K	65–89

105 more models in the catalog do not fit this default profile, or ranked below the first 40 that do.

Example: Runs fully in unified memory on your Apple M2 Max at Q5_K_M, 8K context, about 11–14 tok/s. Weights 18.8 GB plus 0.57 GB of KV cache — 19.9 GB of 24.0 GB usable. Up to 64K context still fits. Weights shown in the verdict use GiB (the same 1024³ counting as GPU memory).