Can the Apple M3 Ultra run gpt-oss 120B?
Yes. gpt-oss 120B needs about 67 GB and the Apple M3 Ultra with 96 GB of unified memory leaves about 72 GB for models. Expect about 99 tok/s, faster than you can read.
ollama run gpt-oss:120bFirst time? Install Ollama, then paste that into a terminal.
gpt-oss 120B needsAbout 67 GB
Apple M3 Ultra has96 or 256 or 512 GB unified memory
Bandwidth819 GB/s
Estimated speed~99 tok/s
By memory size
Each Apple M3 Ultra version
- 96 GB (about 72 GB usable)Runs great ~99 tok/s
- 256 GB (about 192 GB usable)Runs great ~99 tok/s
- 512 GB (about 384 GB usable)Runs great ~99 tok/s
Want it faster?
Speed comes from memory bandwidth: the card reads all 2.9 GB of active weights for every token. On the Apple M3 Ultra, Qwen3 4B runs at about 180 tok/s if you can trade some quality for speed. Cards that run gpt-oss 120B 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