How much RAM does gpt-oss-puzzle-88B need?

gpt-oss-puzzle-88B needs about 54 GB of memory at Q4_K_M quantization with 4K context. The download itself is 50 GB; the extra covers the context cache and runtime overhead.

Can I run gpt-oss-puzzle-88B on a laptop?

Yes — it runs comfortably on a machine with 64GB RAM, for example MacBook Pro M4 Max.

gpt-oss-puzzle-88B: hardware requirements

gpt-oss-puzzle-88B needs about 54 GB of memory at Q4 quantization — a 50 GB download plus room for context and overhead. That fits a machine with 64GB RAM.

Parameters

91B

Tier

Needs a workstation

Licence

See model card

Released

2026-03-26

Will it run on your machine?

8GB RAMNot enough memory 16GB RAMNot enough memory 24GB RAMNot enough memory 32GB RAMNot enough memory 64GB RAMRuns well 128GB RAMRuns well 8GB VRAMNot enough memory 12GB VRAMNot enough memory 16GB VRAMNot enough memory 24GB VRAMNot enough memory 32GB VRAMNot enough memory

Download sizes by quantization

Lower quantization means a smaller file and less memory, at some cost to quality. Q4_K_M is the usual starting point. What is this?

Quant	Download	RAM @ 4K	RAM @ 32K	Source
Published	50 GB	54 GB	56 GB	measured

“Measured” sizes are read from published GGUF files (nvidia/gpt-oss-puzzle-88B). “Estimated” sizes are derived from the parameter count and are typically within a few percent. The RAM columns add the context cache and runtime overhead to the download size.

Model card on HuggingFace →API specs & pricing →Verified 2026-08-09Tracked for 12 months after release