Can the Apple M2 run gpt-oss 20B?
Yes, with 24 GB or more. gpt-oss 20B needs about 16 GB and the Apple M2 with 24 GB of unified memory leaves about 16 GB for models. Expect about 18 tok/s, about reading speed. The 8 and 16 GB versions don't have room.
ollama run gpt-oss:20bFirst time? Install Ollama, then paste that into a terminal.
gpt-oss 20B needsAbout 16 GB
Apple M2 has8 or 16 or 24 GB unified memory
Bandwidth100 GB/s
Estimated speed~18 tok/s
By memory size
Each Apple M2 version
- 8 GB (about 5 GB usable)Too big
- 16 GB (about 11 GB usable)Too big
- 24 GB (about 16 GB usable)Usable ~18 tok/s
Want it faster?
Speed comes from memory bandwidth: the card reads all 2 GB of active weights for every token. Cards that run gpt-oss 20B faster:
How we estimate
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.
Keep going