← Back to all models

llm jp 4 32b a3b thinking

General model by llm-jp · 32.14B parameters

Chip
Unified Memory
Bandwidth 307 GB/s
Available for Models ~21 GB

Everything this chip runs, machine by machine: MacBook Pro 14" M5 Pro, MacBook Pro 16" M5 Pro

Tight Fit

llm jp 4 32b a3b thinking can run but will be tight at 75% of RAM. Consider a lower quantization or closing other apps.

tok/s

Q3_K_M · 18.0 GB · Benchmark needed

Alternative: Q2_K (14.5 GB, ~130 tok/s)

Quantization Options

QuantizationMemorySpeedFits?
Q8_0 36.4 GB~46 tok/s
Q6_K 28.3 GB~60 tok/s
Q5_K_M 24.5 GB~71 tok/s
Q4_K_M 21.2 GB~83 tok/s
Q3_K_M Recommended 18.0 GB~100 tok/s
Q2_K 14.5 GB~130 tok/s

Specifications

Parameters32.1B
Architectureqwen3_moe
Context Length66K
CategoryGeneral
Mixture of Experts 128 experts, 8 active
Formatgguf
Minimum RAM18 GB
Licenseapache-2.0
HuggingFace Downloads33,326

Which Mac Can Run llm jp 4 32b a3b thinking?

Minimum Mac

MacBook Pro 14" — M5 Max 36GB

Q6_K · benchmark needed · 79% RAM

MacQuanttok/sRAM %
Mac Studio · M3 Ultra 96GB Q8_0 38%
Mac Studio · M3 Ultra 256GB Q8_0 14%
Mac Studio · M3 Ultra 512GB Q8_0 7%
MacBook Pro 14" · M5 Max 36GB Q6_K 79%
MacBook Pro 16" · M5 Max 36GB Q6_K 79%
Mac Studio · M1 Ultra 64GB Q8_0 57%
Mac Studio · M1 Ultra 128GB Q8_0 28%
Mac Studio · M2 Ultra 64GB Q8_0 57%
Mac Studio · M2 Ultra 128GB Q8_0 28%
Mac Studio · M2 Ultra 192GB Q8_0 19%
MacBook Pro 14" · M4 Max 36GB Q6_K 79%
MacBook Pro 16" · M4 Max 36GB Q6_K 79%
Mac Studio · M4 Max 36GB Q6_K 79%
MacBook Pro 14" · M5 Max 48GB Q8_0 76%
MacBook Pro 14" · M5 Max 64GB Q8_0 57%
MacBook Pro 14" · M5 Max 128GB Q8_0 28%
MacBook Pro 16" · M5 Max 48GB Q8_0 76%
MacBook Pro 16" · M5 Max 64GB Q8_0 57%
MacBook Pro 16" · M5 Max 128GB Q8_0 28%
MacBook Pro 14" · M4 Max 48GB Q8_0 76%
MacBook Pro 14" · M4 Max 64GB Q8_0 57%
MacBook Pro 14" · M4 Max 128GB Q8_0 28%
MacBook Pro 16" · M4 Max 48GB Q8_0 76%
MacBook Pro 16" · M4 Max 64GB Q8_0 57%
MacBook Pro 16" · M4 Max 128GB Q8_0 28%
Mac Studio · M4 Max 48GB Q8_0 76%
Mac Studio · M4 Max 64GB Q8_0 57%
Mac Studio · M4 Max 128GB Q8_0 28%
MacBook Pro 14" · M3 Max 36GB Q6_K 79%
MacBook Pro 16" · M3 Max 36GB Q6_K 79%
MacBook Pro 14" · M1 Max 64GB Q8_0 57%
MacBook Pro 14" · M2 Max 64GB Q8_0 57%
MacBook Pro 14" · M2 Max 96GB Q8_0 38%
MacBook Pro 14" · M3 Max 48GB Q8_0 76%
MacBook Pro 14" · M3 Max 64GB Q8_0 57%
MacBook Pro 14" · M3 Max 96GB Q8_0 38%
MacBook Pro 14" · M3 Max 128GB Q8_0 28%
MacBook Pro 16" · M1 Max 64GB Q8_0 57%
MacBook Pro 16" · M2 Max 64GB Q8_0 57%
MacBook Pro 16" · M2 Max 96GB Q8_0 38%
MacBook Pro 16" · M3 Max 48GB Q8_0 76%
MacBook Pro 16" · M3 Max 64GB Q8_0 57%
MacBook Pro 16" · M3 Max 96GB Q8_0 38%
MacBook Pro 16" · M3 Max 128GB Q8_0 28%
Mac Studio · M1 Max 64GB Q8_0 57%
Mac Studio · M2 Max 64GB Q8_0 57%
Mac Studio · M2 Max 96GB Q8_0 38%
MacBook Pro 14" · M5 Pro 48GB Q8_0 76%
MacBook Pro 14" · M5 Pro 64GB Q8_0 57%
MacBook Pro 16" · M5 Pro 48GB Q8_0 76%
MacBook Pro 16" · M5 Pro 64GB Q8_0 57%
MacBook Pro 14" · M4 Pro 48GB Q8_0 76%
MacBook Pro 16" · M4 Pro 48GB Q8_0 76%
Mac mini · M4 Pro 48GB Q8_0 76%
Mac mini · M4 Pro 64GB Q8_0 57%
MacBook Pro 14" · M3 Pro 36GB Q6_K 79%
MacBook Pro 16" · M3 Pro 36GB Q6_K 79%
MacBook Pro 14" · M5 Pro 24GB Q3_K_M 75%
MacBook Pro 16" · M5 Pro 24GB Q3_K_M 75%
MacBook Pro 14" · M1 Max 32GB Q5_K_M 76%
MacBook Pro 14" · M2 Max 32GB Q5_K_M 76%
MacBook Pro 16" · M1 Max 32GB Q5_K_M 76%
MacBook Pro 16" · M2 Max 32GB Q5_K_M 76%
Mac Studio · M1 Max 32GB Q5_K_M 76%
Mac Studio · M2 Max 32GB Q5_K_M 76%
MacBook Pro 14" · M4 Pro 24GB Q3_K_M 75%
MacBook Pro 16" · M4 Pro 24GB Q3_K_M 75%
Mac mini · M4 Pro 24GB Q3_K_M 75%
MacBook Pro 14" · M3 Pro 18GB Q2_K 81%
MacBook Pro 16" · M3 Pro 18GB Q2_K 81%
MacBook Air · M5 24GB Q3_K_M 75%
MacBook Pro 14" · M5 24GB Q3_K_M 75%
MacBook Pro 14" · M1 Pro 32GB Q5_K_M 76%
MacBook Pro 14" · M2 Pro 32GB Q5_K_M 76%
MacBook Pro 16" · M1 Pro 32GB Q5_K_M 76%
MacBook Pro 16" · M2 Pro 32GB Q5_K_M 76%
Mac mini · M2 Pro 32GB Q5_K_M 76%
MacBook Air · M4 24GB Q3_K_M 75%
MacBook Pro 14" · M4 24GB Q3_K_M 75%
Mac mini · M4 24GB Q3_K_M 75%
iMac · M4 24GB Q3_K_M 75%
MacBook Air · M5 32GB Q5_K_M 76%
MacBook Pro 14" · M5 32GB Q5_K_M 76%
MacBook Air · M2 24GB Q3_K_M 75%
MacBook Air · M3 24GB Q3_K_M 75%
MacBook Pro 14" · M3 24GB Q3_K_M 75%
Mac mini · M2 24GB Q3_K_M 75%
iMac · M3 24GB Q3_K_M 75%
MacBook Air · M4 32GB Q5_K_M 76%
MacBook Pro 14" · M4 32GB Q5_K_M 76%
Mac mini · M4 32GB Q5_K_M 76%
iMac · M4 32GB Q5_K_M 76%
MacBook Pro 14" · M1 Pro 16GB 91%
MacBook Pro 14" · M2 Pro 16GB 91%
MacBook Pro 16" · M1 Pro 16GB 91%
MacBook Pro 16" · M2 Pro 16GB 91%
Mac mini · M2 Pro 16GB 91%
MacBook Air · M5 16GB 91%
MacBook Pro 14" · M5 16GB 91%
MacBook Air · M4 16GB 91%
MacBook Pro 14" · M4 16GB 91%
Mac mini · M4 16GB 91%
iMac · M4 16GB 91%
MacBook Air · M2 8GB 181%
MacBook Air · M2 16GB 91%
MacBook Air · M3 8GB 181%
MacBook Air · M3 16GB 91%
MacBook Pro 14" · M3 8GB 181%
MacBook Pro 14" · M3 16GB 91%
Mac mini · M2 8GB 181%
Mac mini · M2 16GB 91%
iMac · M3 8GB 181%
iMac · M3 16GB 91%
MacBook Air · M1 8GB 181%
MacBook Air · M1 16GB 91%
Mac mini · M1 8GB 181%
Mac mini · M1 16GB 91%
iMac · M1 8GB 181%
iMac · M1 16GB 91%

Run llm jp 4 32b a3b thinking locally on your Mac

ToolPiper downloads, manages, and runs models with one click. Apple Silicon optimized.

Get ToolPiper — Free