RunsOnMy ModelsGPUs & MacsAPI pricesVRAM calculatorNew

What AI models run on the Mac M1 Pro 32GB?

The Mac M1 Pro 32GB has 32 GB of unified memory (200 GB/s). About 21.4 GB is usable for a model. 202 of 285 open models we track fit at 8K context.

202models fit (8K ctx)
196still fit at 32K ctx
49Blargest that fits
200 GB/smemory bandwidth

Models that fit, largest first

ModelParamsBest quant (8K)Needs32K contextMax tok/s (est.)
Kimi Linear 48B A3B Instruct49B (3.0B act.)Q2_K17.3 GBQ2_K187details
Nous Hermes 2 Mixtral 8x7B DPO47B (13B act.)Q2_K17.6 GBQ2_K42details
falcon 40b42BQ2_K16.1 GBQ2_K13details
flint40B (1.3B act.)Q3_K_M19.0 GBQ2_K327details
Maxwell V1 Flash37B (2.3B act.)Q3_K_M17.9 GBQ3_K_M180details
Hermes 4.3 36B36BQ3_K_M19.1 GBQ2_K11details
Seed OSS 36B Instruct36BQ3_K_M19.1 GBQ2_K11details
Qwen3.6 35B A3B36B (3.0B act.)Q3_K_M17.0 GBQ3_K_M140details
Ornith 1.5 35B A3B36B (3.0B act.)Q3_K_M16.6 GBQ3_K_M144details
Qwen3.5 35B A3B36B (3.0B act.)Q3_K_M16.3 GBQ3_K_M147details
Axion 1.0 35B35B (3.9B act.)Q3_K_M16.6 GBQ3_K_M107details
Qwen AgentWorld 35B A3B35B (3.0B act.)Q3_K_M16.6 GBQ3_K_M139details
Yi 34B 200K34BQ3_K_M17.9 GBno12details
Yi 34B34BQ3_K_M17.9 GBno12details
Yi 34B Chat34BQ3_K_M17.9 GBno12details
Yi 1.5 34B Chat34BQ3_K_M18.1 GBQ2_K12details
Nous Hermes 2 Yi 34B34BQ3_K_M18.4 GBQ2_K12details
DeepSeek R1 Distill Qwen 32B33BQ3_K_M17.5 GBQ2_K13details
Qwen3 32B33BQ3_K_M17.6 GBQ2_K12details
GLM Z1 32B 041433BQ3_K_M16.0 GBQ3_K_M13details
GLM 4 32B 041433BQ3_K_M16.0 GBQ3_K_M13details
granite 4.0 h small32B (8.8B act.)Q4_K_M18.9 GBQ4_K_M38details
gemma 4 31B it31BQ3_K_M16.7 GBQ2_K14details
GLM 4.7 Flash31B (4.6B act.)Q4_K_M18.2 GBQ3_K_M74details
Qwen3 Coder 30B A3B Instruct31B (3.0B act.)Q4_K_M18.7 GBQ3_K_M110details
North Mini Code 1.030B (3.3B act.)Q4_K_M19.3 GBQ3_K_M96details
Hy Memory 30B A3B30B (3.0B act.)Q4_K_M18.5 GBQ3_K_M109details
Muse Glimmer 30B30BQ5_K_M18.8 GBQ5_K_M10details
granite 4.2 30b29BQ4_K_M19.1 GBQ2_K11details
Qwen3.8 27B28BQ4_K_M16.8 GBQ4_K_M12details
Qwen3.6 27B28BQ4_K_M17.0 GBQ4_K_M12details
Qwen3.5 27B28BQ4_K_M17.0 GBQ4_K_M12details
gemma 3 27b it27BQ4_K_M17.3 GBQ4_K_M12details
Arete 27B27BQ4_K_M17.4 GBQ4_K_M12details
MiniMax H3 Image PE27BQ5_K_M19.3 GBQ4_K_M10details
Trinity Mini26B (2.0B act.)Q5_K_M18.4 GBQ5_K_M143details
GLiDE 27B26BQ5_K_M18.7 GBQ4_K_M11details
diffusiongemma 26B A4B it26B (4.0B act.)Q5_K_M19.2 GBQ4_K_M67details
gemma 4 26B A4B it26B (4.0B act.)Q4_K_M17.2 GBQ4_K_M76details
LFM2.5 24B A2B24B (2.0B act.)Q6_K19.1 GBQ5_K_M122details
ERNIE 4.5 21B A3B PT22BQ6_K17.8 GBQ6_K11details
ERNIE 4.5 21B A3B Thinking22BQ6_K17.8 GBQ6_K11details
gpt oss safeguard 20b22B (4.8B act.)F1613.8 GBF1665details
gpt oss 20b21B (4.2B act.)F1613.8 GBF1672details
Groot 19B A3B 26102319B (3.0B act.)Q6_K16.7 GBQ6_K77details
Kimi VL A3B Thinking 250616B (3.0B act.)Q8_016.8 GBQ8_064details
Kimi VL A3B Instruct16B (3.0B act.)Q8_016.8 GBQ8_064details
Kimi VL A3B Thinking16B (3.0B act.)Q8_017.2 GBQ8_063details
deepseek moe 16b base16B (2.8B act.)Q8_018.6 GBQ5_K_M66details
LLaDA2.1 mini16B (1.4B act.)Q8_017.1 GBQ8_0131details
Moonlight 16B A3B Instruct16B (3.0B act.)Q8_016.8 GBQ8_063details
DeepSeek Coder V2 Lite Instruct16B (2.7B act.)Q8_016.4 GBQ8_071details
DeepSeek V2 Lite16B (2.7B act.)Q8_016.4 GBQ8_071details
DeepSeek V2 Lite Chat16B (2.7B act.)Q8_016.4 GBQ8_071details
Qwen2.5 Coder 14B Instruct15BQ8_016.8 GBQ6_K13details
DeepSeek R1 Distill Qwen 14B15BQ8_016.8 GBQ6_K13details
Qwen3 14B15BQ8_016.6 GBQ6_K13details
phi 415BQ8_016.7 GBQ6_K13details
gemma 3 12b it12BQ8_013.4 GBQ8_016details
Llama Guard 4 12B12BQ8_014.2 GBQ8_016details
Llama 3.2 11B Vision Instruct11BQ8_011.6 GBQ8_019details
Llama 3.2 11B Vision11BQ8_012.4 GBQ8_018details
GLM 4.6V Flash10BF1618.5 GBQ8_011details
GLM 4.1V 9B Thinking10BF1618.5 GBQ8_011details
Step3 VL 10B10BQ8_09.9 GBQ8_023details
SenseNova Loop 9B Preview9.7BF1617.8 GBF1611details
Qwen3.5 9B9.7BF1617.8 GBF1611details
Ornith 1.5 9B9.7BF1617.8 GBF1611details
Qwen3.5 9B Base9.7BF1617.8 GBF1611details
MiMo V2.6 Distill Qwen 9B9.4BF1617.8 GBF1611details
glm 4 9b9.4BQ8_015.0 GBno20details
internlm3 8b instruct8.8BF1617.4 GBF1611details
granite 4.2 8b8.8BF1618.2 GBQ8_011details
granite 4.1 8b8.8BF1618.2 GBQ8_011details
Qwen3 VL 8B Instruct8.8BF1617.1 GBQ8_012details
Molmo2 8B8.7BF1617.1 GBQ8_012details
LFM2.5 8B A1B8.5B (1.0B act.)Q8_09.1 GBQ8_0188details
Qwen2.5 VL 7B Instruct8.3BF1615.3 GBF1613details
UI TARS 1.5 7B8.3BF1615.3 GBF1613details
olmOCR 2 7B 10258.3BF1615.3 GBF1613details
olmOCR 7B 0225 preview8.3BF1616.6 GBF1612details
UI TARS 7B DPO8.3BF1615.3 GBF1613details
Qwen3 8B8.2BF1617.1 GBQ8_012details
DeepSeek R1 0528 Qwen3 8B8.2BF1617.1 GBQ8_012details
granite 3.3 8b instruct8.2BF1617.0 GBQ8_012details
granite 3.1 8b instruct8.2BF1617.0 GBQ8_012details
granite 3.2 8b instruct8.2BF1617.0 GBQ8_012details
Apertus 8B Instruct 25098.1BF1616.7 GBQ8_012details
Llama 3.1 8B Instruct8.0BF1616.6 GBQ8_012details
Llama 3.1 8B8.0BF1616.6 GBQ8_012details
Hermes 3 Llama 3.1 8B8.0BF1616.6 GBQ8_012details
DeepSeek R1 Distill Llama 8B8.0BQ8_09.6 GBQ8_023details
Meta Llama 3 8B Instruct8.0BF1616.6 GBQ8_012details
Meta Llama Guard 2 8B8.0BF1616.6 GBQ8_012details
Meta Llama 3.1 8B Instruct8.0BF1616.6 GBQ8_012details
Ling 3.0 tiny7.9B (1.4B act.)F1615.6 GBF1672details
MiMo 7B RL7.8BF1616.4 GBQ8_013details
internlm2 5 7b chat7.7BF1616.0 GBF1613details
internlm2 chat 7b7.7BF1616.0 GBF1613details
Qwen2.5 7B Instruct7.6BF1615.3 GBF1613details
Qwen2.5 Coder 7B Instruct7.6BF1615.3 GBF1613details
DeepSeek R1 Distill Qwen 7B7.6BF1615.3 GBF1613details
Falcon H1R 7B7.6BF1615.1 GBF1613details
Olmo 3 7B Instruct7.3BF1616.7 GBQ8_014details
Olmo 3 1025 7B7.3BF1616.7 GBQ8_014details
Olmo 3 7B Think7.3BF1616.7 GBQ8_014details
Mistral 7B v0.17.2BF1614.9 GBF1614details
Mistral 7B Instruct v0.27.2BF1614.9 GBF1614details
Mistral 7B Instruct v0.17.2BF1614.9 GBF1614details
falcon 7b7.2BF1618.4 GBno14details
falcon 7b instruct7.2BF1618.4 GBno14details
granite 4.0 h tiny6.9B (1.5B act.)F1613.6 GBF1668details
OLMoE 1B 7B 09246.9B (1.3B act.)F1614.4 GBF1678details
deepseek coder 7b instruct v1.56.9BF1617.2 GBQ3_K_M14details
deepseek coder 6.7b instruct6.7BF1617.0 GBQ2_K15details
deepseek coder 6.7b base6.7BF1617.0 GBQ2_K15details
Llama 2 7b hf6.7BF1617.0 GBQ2_K15details
Yi 6B 200K6.1BF1612.3 GBF1616details
Yi 6B6.1BF1612.3 GBF1616details
Molmo2 4B4.9BF1610.0 GBF1623details
WeVisOne VL 4B4.8BF1610.0 GBF1623details
Mage VL Cowork4.7BF1610.0 GBF1623details
Qianfan OCR4.7BF1610.1 GBF1623details
Qwen3.5 4B4.7BF169.0 GBF1624details
S5 nano4.7BF169.2 GBF1623details
LightOnOCR 3 4B4.5BF169.0 GBF1624details
SingVL S1 4B4.5BF169.0 GBF1624details
Intern Decision 4B4.5BF169.0 GBF1624details
Qwen3 VL 4B Instruct4.4BF169.3 GBF1625details
gemma 3 4b it4.3BF168.4 GBF1626details
medgemma 1.5 4b it4.3BF168.4 GBF1626details
medgemma 4b it4.3BF168.4 GBF1626details
gemma 3 4b pt4.3BF169.2 GBF1623details
Phi 3.5 vision instruct4.1BF1611.2 GBQ8_024details
Qwen3 4B4.0BF169.3 GBF1625details
Qwen3 4B Instruct 25074.0BF169.3 GBF1625details
granite vision 4.1 4b4.0BF167.6 GBF1629details
Phi 4 mini instruct3.8BF168.9 GBF1626details
Phi 3 mini 4k instruct3.8BF1610.6 GBQ8_026details
Phi 3.5 mini instruct3.8BF1610.6 GBQ8_026details
Phi 3 mini 128k instruct3.8BF1610.6 GBQ8_026details
Qwen2.5 VL 3B Instruct3.8BF166.7 GBF1632details
granite 4.2 3b3.7BF168.0 GBF1627details
granite 4.1 3b llmon preview3.7BF168.0 GBF1627details
Raon 1 nano3.5BF167.0 GBF1632details
granite 4.1 3b3.4BF167.6 GBF1629details
granite 4.0 micro3.4BF167.6 GBF1629details
DeepSeek OCR 23.4B (1.2B act.)F166.6 GBF1697details
deepseek vl2 tiny3.4B (1.2B act.)F167.4 GBF1685details
Unlimited OCR 23.3B (1.1B act.)F166.6 GBF1699details
Unlimited OCR3.3B (1.1B act.)F166.6 GBF16100details
DeepSeek OCR3.3B (1.1B act.)F166.6 GBF16100details
granite 3.1 3b a800m instruct3.3B (883M act.)F167.1 GBF16113details
Llama 3.2 3B Instruct3.2BF1613.5 GBF1616details
Llama 3.2 3B3.2BF167.5 GBF1631details
Hermes 3 Llama 3.2 3B3.2BF167.5 GBF1631details
d1 3B3.1BF1610.8 GBF1619details
LFM2.5 VL 3B3.1BF1610.8 GBF1619details
Qwen2.5 3B Instruct3.1BF166.7 GBF1632details
SmolLM3 3B3.1BF166.9 GBF1632details
SmolLM3 3B Base3.1BF166.9 GBF1633details
phi 22.8BF168.2 GBF1636details
LFM2.5 2.6B2.7BF1610.8 GBF1619details
granite 3.3 2b instruct2.5BF165.8 GBF1639details
MiniCPM5 2B2.5BF165.7 GBF1640details
WeVisOne VL 2B2.4BF165.4 GBF1649details
Qwen3.5 2B2.3BF164.5 GBF1653details
SmolVLM Instruct2.2BF165.4 GBF1655details
Qwen3 1.7B2.0BF164.8 GBF1658details
Qwen3.8 27B DSpine b162.0BF164.8 GBF1649details
d1 2B1.9BF167.2 GBF1629details
LFM3 VL 2B1.9BF167.2 GBF1629details
DeepSeek R1 Distill Qwen 1.5B1.8BF164.2 GBF1656details
LFM3 1.7B1.7BF167.2 GBF1629details
LFM3 1.7B Base1.7BF164.0 GBF1658details
SmolLM2 1.7B Instruct1.7BF165.2 GBF1658details
SmolLM2 1.7B1.7BF165.2 GBF1658details
SmolLM 1.7B1.7BF165.2 GBF1658details
granite 4.0 1b1.6BF164.3 GBF1661details
Qwen2.5 1.5B Instruct1.5BF163.8 GBF1665details
OLMo 2 0425 1B1.5BF164.4 GBF1667details
deepseek coder 1.3b instruct1.3BF164.5 GBF1674details
Llama 3.2 1B1.2BF163.2 GBF1681details
Llama 3.2 1B Instruct1.2BF165.5 GBF1640details
LFM2.5 1.2B Instruct1.2BF165.0 GBF1643details
HunyuanOCR1.1BF163.5 GBF1689details
PaddleOCR VL 1.6959MF162.5 GBF16104details
Qwen3.5 0.8B873MF162.4 GBF16132details
LightOnOCR 3 0.8B853MF162.4 GBF16132details
FastVLM 0.5B759MF162.2 GBF16132details
Qwen3 0.6B752MF162.7 GBF16167details
GOT OCR2 0716MF162.8 GBF16140details
Nemotron 3.5 Parse631MF161.9 GBF16153details
GOT OCR 2.0 hf561MF162.5 GBF16178details
SmolVLM2 500M Video Instruct507MF161.6 GBF16244details
SmolVLM 500M Instruct507MF161.6 GBF16244details
Qwen2.5 0.5B Instruct494MF161.7 GBF16202details
LFM2.5 VL 450M449MF161.9 GBF16141details
SmolLM2 360M Instruct362MF161.5 GBF16276details
SmolLM2 360M362MF161.5 GBF16276details
LFM2.5 350M354MF161.9 GBF16141details
LFM2 350M354MF161.4 GBF16282details

Just too big at 4-bit

These need up to 35% more memory than the Mac M1 Pro 32GB has at Q4_K_M and 8K context. A smaller quant or shorter context may work.

ModelQ4_K_M needsShort by
Qwen3.5 35B A3B21.5 GB0.1 GB
Yi 34B21.6 GB0.2 GB
Yi 34B 200K21.6 GB0.2 GB
Yi 34B Chat21.6 GB0.2 GB
Yi 1.5 34B Chat21.6 GB0.2 GB
Qwen AgentWorld 35B A3B21.6 GB0.2 GB
Maxwell V1 Flash22.1 GB0.6 GB
Qwen3.6 35B A3B22.1 GB0.7 GB
Nous Hermes 2 Yi 34B22.1 GB0.7 GB
Hermes 4.3 36B23.0 GB1.5 GB
Seed OSS 36B Instruct23.0 GB1.5 GB
flint23.6 GB2.1 GB
falcon 40b25.2 GB3.8 GB
Nous Hermes 2 Mixtral 8x7B DPO28.0 GB6.5 GB

Questions

What is the biggest model that runs on the Mac M1 Pro 32GB?

Of the models we track, the largest that fits at 8K context is Kimi Linear 48B A3B Instruct (49B) at Q2_K, needing about 17.3 GB.

How much of the Mac M1 Pro 32GB's memory can a model use?

About 21 GB of 32 GB. macOS lets the GPU use about 67% of unified memory by default; this can be raised with the iogpu.wired_limit_mb setting.

Spec source: Apple specifications. Model data: Hugging Face Hub API. Speed is a bandwidth ceiling (bandwidth ÷ bytes read per token), not a benchmark. Methodology · Report an error.