Can the Apple M1 run Gemma 3 12B?
Yes, with 16 GB or more. Gemma 3 12B needs about 10 GB and the Apple M1 with 16 GB of unified memory leaves about 11 GB for models. Expect about 5 tok/s, slow, but it works. The 8 GB version doesn't have room.
ollama run gemma3:12bFirst time? Install Ollama, then paste that into a terminal.
Each Apple M1 version
- 8 GB (about 5 GB usable)Too big
- 16 GB (about 11 GB usable)Slow ~5 tok/s
Speed comes from memory bandwidth: the card reads all 8.1 GB of active weights for every token. On the Apple M1, Qwen3 8B runs at about 7 tok/s if you can trade some quality for speed. Cards that run Gemma 3 12B faster:
Speed is roughly memory bandwidth divided by the weights read per token, calibrated on our own rack. Sizes are Ollama's default 4-bit downloads plus about 1.2 GB for context. Real results are usually within about a third. Measured it yourself? Send us the number.