1bit.MONSTERDocs GitHub ↗

GPT-OSS — Open MoE

Open-weight 20B MoE. Both the base and safeguard variants are pre-compiled for NPU with dedicated expert-dispatch xclbins.

Models

Model Params 1BP Size Backend(s) Perf
GPT-OSS-20B 20B NPU / CPU
GPT-OSS-Safeguard-20B 20B NPU / CPU

Notes

See also: full model support detail · benchmarks SSOT · all families