text model · gpt-oss · macOS
Can I run gpt-oss 120B on Apple M3 Pro (18GB)?
No. gpt-oss 120B needs ~62.4 GB even at Q4_K_M, but Apple M3 Pro (18GB) only has ~12 GB usable.
Needs ~62.4 GB even at Q4_K_M, but only ~12 GB is usable.
- Q4_K_M needed
- ~62.4 GB
- Usable on device
- ~12 GB
- Device memory
- 18 GB
- Parameters
- 117B (MoE, 5.1B active)
- Q4_K_M size
- 59.03 GB
- Context
- 128k
- Ollama tag
- gpt-oss:120b
- Memory
- 18 GB unified
- Usable for weights
- ~12 GB
- Best runtime
- Ollama (llama.cpp Metal backend) / MLX
What you can run instead
Run gpt-oss 120B on other hardware
FAQ
Can Apple M3 Pro (18GB) run gpt-oss 120B?
No. gpt-oss 120B needs ~62.4 GB even at Q4_K_M, but Apple M3 Pro (18GB) only has ~12 GB usable.
How much memory does gpt-oss 120B need?
Apple M3 Pro (18GB) does not have enough memory. At Q4_K_M the weights are ~59.03 GB; with KV cache and runtime overhead, budget ~62.4 GB at a 4k context. It is a Mixture-of-Experts model (117B total / 5.1B active), so all experts must stay in memory; memory tracks total params, not active params.
What is the best tool to run gpt-oss 120B on macOS?
LM Studio for a simple setup; mlx-lm for the most speed. vLLM is NOT a Mac tool, it is a CUDA/Linux serving engine. Unified memory is not a fixed VRAM slice; ~70% is usable for weights.
Sources
Memory figures are estimates. See methodology.