model · OpenAI
gpt-oss-20b
gpt-oss-20b is OpenAI's smaller open-weight model — a mixture-of-experts design with 21B total parameters but only ~3.6B active per token, so it runs fast and fits in about 16 GB of memory. It ships natively in the 4-bit MXFP4 format, has a 128K context window, and offers strong tool-use and reasoning for its size, under an Apache-2.0 licence. It is the sweet spot: a genuinely capable model you can run privately on a normal laptop or a single consumer GPU.
What it can help with
- laptop inference
- low vram
- moe architecture
- 4-bit quantization
- open weight
- apache licence
- offline