MacBook Air M3 16GB

Apple · M3 · 16GB Unified Memory · Can run 19 models

Manufacturer Apple
Unified Memory 16 GB
Chip M3
CPU Cores 8
GPU Cores 10
Neural Engine Cores 16
Memory Bandwidth 100 GB/s
MSRP $1,099
Released Mar 8, 2024

AI Notes

The MacBook Air M3 16GB is a fanless, portable option for running local AI models. With 16GB of unified memory, it handles 7B models well and can run 13B models with quantization. The 100 GB/s memory bandwidth is lower than M4, resulting in slightly slower token generation, but it remains capable for on-the-go AI work.

Compatible Models

Model Parameters Best Quant VRAM Used Fit
Llama 3.2 1B 1B Q8_0 3 GB Runs
Gemma 2 2B 2B Q8_0 4 GB Runs
Llama 3.2 3B 3B Q8_0 5 GB Runs
Phi-3 Mini 3.8B 3.8B Q8_0 5.8 GB Runs
DeepSeek R1 7B 7B Q8_0 9 GB Runs
Mistral 7B 7B Q8_0 9 GB Runs
Qwen 2.5 7B 7B Q8_0 9 GB Runs
Qwen 2.5 Coder 7B 7B Q8_0 9 GB Runs
Llama 3.1 8B 8B Q8_0 10 GB Runs
Gemma 2 9B 9B Q8_0 11 GB Runs
DeepSeek R1 14B 14B Q4_K_M 9.9 GB Runs
Phi-4 14B 14B Q4_K_M 9.9 GB Runs
Qwen 2.5 14B 14B Q4_K_M 9.9 GB Runs
Codestral 22B 22B Q4_K_M 14.7 GB Runs (tight)
StarCoder2 15B 15B Q8_0 17 GB CPU Offload
Gemma 2 27B 27B Q4_K_M 17.7 GB CPU Offload
DeepSeek R1 32B 32B Q4_K_M 20.7 GB CPU Offload
Qwen 2.5 32B 32B Q4_K_M 20.7 GB CPU Offload
Command R 35B 35B Q4_K_M 22.5 GB CPU Offload
6 model(s) are too large for this hardware.