Hardware

Apple M1 Ultra

Apple M1 Ultra has 64 / 128 GB Unified and 800 GB/s. See which open models fit, at which quant, and roughly how fast. Figures below use the default memory (64 GB). Updated 2026-10-17.

Memory64 / 128 GB Unified

Bandwidth800 GB/s

Kindmac

Year2022

800 GB/s. 64GB or 128GB. Source.

RSS for new models on this device

Models that fit

Model	Verdict	Quant	Context	tok/s
Llama 3.3 70B Instruct 70.6B	Tight fit	UD-Q4_K_XL	8K	9.7–13
DeepSeek R1 Distill Llama 70B 70.6B	Tight fit	UD-Q4_K_XL	8K	9.7–13
Hermes 4 70B 70.6B	Tight fit	UD-Q4_K_XL	8K	9.7–13
Qwen3 Next 80B A3B Instruct 79.7B	Tight fit	UD-Q4_K_XL	8K	173–236
Qwen3 Next 80B A3B Thinking 79.7B	Tight fit	UD-Q4_K_XL	8K	174–237
Qwen3 Coder Next 79.7B	Tight fit	Q4_K_M	8K	49–66
Mistral Small 4 119B 2603 119B	Tight fit	UD-IQ3_S	8K	53–73
Qwen3.5 122B A10B MTP 125B	Tight fit	UD-IQ3_XXS	8K	109–149
Llama 4 Scout 17B 16E Instruct 108B	Tight fit	Q3_K_S	8K	38–52
Qwen2.5 32B Instruct 32.8B	Runs	Q8_0	8K	12–16
Qwen2.5 Coder 32B Instruct 32.8B	Runs	Q8_0	8K	12–16
DeepSeek R1 Distill Qwen 32B 32.8B	Runs	Q8_0	8K	12–16
Qwen3 32B 32.8B	Runs	UD-Q8_K_XL	8K	11–14
Qwen3 VL 32B Instruct 32.8B	Runs	UD-Q8_K_XL	8K	11–14
GLM 4 32B 0414 32.6B	Runs	UD-Q8_K_XL	8K	11–15
GLM 4.5 Air 110B	Tight fit	UD-Q2_K_XL	8K	38–52
gemma 4 31B it 30.7B	Runs	UD-Q8_K_XL	8K	10–14
GLM 4.6V 107B	Tight fit	UD-Q2_K_XL	8K	39–53
Qwen3.8 27B 27.3B	Runs great	UD-Q8_K_L	8K	15–21
gemma 3 27b it 27.0B	Runs	UD-Q8_K_XL	8K	12–17
medgemma 27b text it 27.0B	Runs	UD-Q8_K_XL	8K	12–17
Qwen3.6 27B 26.9B	Runs	UD-Q8_K_XL	8K	12–17
Qwen3.5 27B 26.9B	Runs	UD-Q8_K_XL	8K	12–17
Kwaipilot KAT Coder V2.5 Dev 34.7B	Runs	Q8_0	8K	64–87
Qwen3.8 35B A3B Distill 35.5B	Runs	Q8_0	8K	131–178
Qwen3.6 35B A3B 34.7B	Runs	UD-Q8_K_XL	8K	126–172
Qwen3.5 35B A3B 34.7B	Tight fit	UD-Q8_K_XL	8K	100–137
NVIDIA Nemotron 3.5 Lightning 30B A3B 32.9B	Runs	UD-Q8_K_XL	8K	111–152
NVIDIA Nemotron 3 Nano Omni 30B A3B Reasoning 31.6B	Runs	UD-Q8_K_XL	8K	116–158
Nemotron 3 Nano 30B A3B 31.6B	Runs	UD-Q8_K_XL	8K	103–140
GLM 4.7 Flash 29.9B	Runs	UD-Q8_K_XL	8K	55–75
Qwen3 Coder 30B A3B Instruct 30.5B	Runs	UD-Q8_K_XL	8K	101–138
Qwen3 30B A3B Instruct 2507 30.5B	Runs	UD-Q8_K_XL	8K	101–138
Qwen3 VL 30B A3B Instruct 30.5B	Runs	UD-Q8_K_XL	8K	101–138
Qwen3 30B A3B Thinking 2507 30.5B	Runs	UD-Q8_K_XL	8K	101–138
gemma 4 26B A4B it 25.2B	Runs great	UD-Q8_K_XL	8K	69–94
LFM2.5 24B A2B 24.0B	Tight fit	BF16	8K	94–128
GLM 4.7 Flash REAP 23B A3B 23.0B	Tight fit	BF16	8K	68–93
gpt oss 20b 20.9B	Runs great	UD-Q8_K_XL	8K	109–149
Qwen2.5 14B Instruct 14.8B	Runs	F16	8K	14–19

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

Example: Tight fit fully in unified memory on your Apple M1 Ultra at UD-Q4_K_XL, 8K context, about 9.7–13 tok/s. Weights 39.7 GB plus 2.50 GB of KV cache — 42.9 GB of 48.0 GB usable. Up to 16K context still fits. Weights shown in the verdict use GiB (the same 1024³ counting as GPU memory).