marcelone

/

Jinx-Qwen3-30B-A3B-Thinking-2507-gguf

like 1

Greetings! Maybe off topic, but you had the MXFP4 model.

https://huggingface.co/marcelone/jinx-gpt-oss-20b-gguf/resolve/main/jinx-gpt-oss-20b-gguf_mxfp4_moe.gguf

It can't be downloaded anymore, could you upload it again? or upload to a file sharing site?

it is very good and consumes 22 gigabytes less memory than the other Q8_0.

marcelone/jinx-gpt-oss-20b-gguf