OpenAI

OpenAI

OpenAI/gpt-oss-20b

Quantizations

Quant	Quantized by	Size	Decode	Prefill	Score	Actions
mxfp4	MLX Community MLX Community	10.4 GB	N/A	N/A	N/A	Run
mxfp4	MLX Community MLX Community	11.2 GB	94.2 tok/s	1,267.9 tok/s	Runs well	Run
q8_0	Unsloth Unsloth	11.3 GB	88.5 tok/s	1,153.2 tok/s	Runs well	Run
4bit	OpenAI OpenAI	12.8 GB	N/A	N/A	N/A	Run
8bit	LM Studio LM Studio	20.7 GB	69.8 tok/s	1,302.4 tok/s	Runs ok	Run

Device Comparison

Results include trials with 4,096 input tokens and 1,024 output tokens only.