RunsLocalGPU checkerLocal AI toolsv1.8 · data Oct 14, 2026

All models / gpt-oss-20b

What GPU do I need to run gpt-oss-20b?

gpt-oss-20b has 21B parameters (3.6B active per token, mixture-of-experts). At Q4_K_M with 8K context it needs about 13.6 GB of memory. The smallest graphics card that runs it fully on the GPU is about 16 GB, such as the RTX 4060 Ti 16GB. Because only 3.6B parameters are active, it still runs usably with offload to system RAM on cards like the RTX 3080 Ti.

Check your own hardware

Memory by quantization (8K context)

Hardware that runs it fully in fast memory

Runs with offload to system RAM (slower)

RTX 3080 Ti ~84 tok/s · RTX 3080 12GB ~84 tok/s · RTX 5070 ~77 tok/s · RTX 4070 Ti ~69 tok/s · RTX 4070 Super ~69 tok/s · RTX 4070 ~69 tok/s · Arc B580 ~66 tok/s · RX 7700 XT ~64 tok/s · RTX 2080 Ti ~59 tok/s · RTX 3060 12GB ~59 tok/s · RTX 3080 10GB ~51 tok/s · RTX 5060 Ti 8GB ~36 tok/s · RTX 5060 ~36 tok/s · RTX 3070 ~36 tok/s · RTX 3060 Ti ~36 tok/s · RTX 5050 ~33 tok/s · RTX 4060 Ti 8GB ~33 tok/s · RX 7600 ~33 tok/s · RTX 4060 ~32 tok/s

Assumes 32 GB of DDR5 system RAM for offload. Estimates only; real speeds usually land within about 20%.