The compatibility matrix

37 models × 16 machines = 592 verdicts — 560 of them run. Every cell is the same GQA-aware fit math as the CLI; click any cell for the full verdict.

✓ comfortable✓ tight~ offload/CPU spill✗ won't fit
modelApple M1
8GB
Apple M1 Pro
16GB
Apple M2
16GB
Apple M3 Pro
18GB
Apple M2 Pro
32GB
Apple M4 Pro
48GB
Apple M3 Max
36GB
Apple M2 Max
64GB
Apple M4 Max
128GB
Apple M2 Ultra
192GB
NVIDIA GeForce RTX 3060
12GB
NVIDIA GeForce RTX 4070
12GB
NVIDIA GeForce RTX 4090
24GB
NVIDIA GeForce RTX 5080
16GB
NVIDIA GeForce RTX 5090
32GB
AMD Radeon RX 7900 XTX
24GB
DeepSeek-R1 (distill)
Llama 3.1~
Qwen2.5
Qwen2.5-Coder
Qwen3
Gemma 4~~~
Llama 3.2
Llama 3.3~~~~
Mistral 7B~
Mistral Small 3~~~~~
Gemma 3
Mistral Nemo~~
Gemma 2
Phi-4~~
Nomic Embed Text
Mixtral 8x7B~~~~~~~
Codestral~~~~~
Phi-4-mini~
Llama 3.2 Vision
mxbai-embed-large
Phi-3.5-mini~
SmolLM2~
StarCoder2
Code Llama~~
Yi 1.5~
Command R~~~~~~~
LLaVA~~
TinyLlama
Granite 3.1 8B~
Falcon3 7B~
OLMo 2 7B~~
Yi-Coder 9B~
Nemotron-Mini 4B~
QwQ 32B~~~~~
Qwen3 Coder 30B-A3B~~~~
Granite 3.3 8B~
SmolLM3 3B

Your machine isn't a column? Check your exact rig → · Rankings per machine: Apple M1 · Apple M1 Pro · Apple M2 · Apple M3 Pro