Model request Qwen3-Coder-Next-REAM 60B mxfp4_moe
#3
by Ohjaaja - opened
Been following your work in the past and i like your quantized models. I am interested in Qwen3-Coder-Next model. Tried 40B and 48B reap variants in my rig, 32gb ram and 12Gb vram. They perform quite well running llama.cpp with partial offloading. Some 20-30t/s. 60B ream might be streching but if it works i might get aother gpu.
Hi. I actually create new model reques for newer version of this.
Ohjaaja changed discussion status to closed