Qwen 3.8 Flash Next (180B-A6B, IQ2_XXS/Q2_K, DS4)

An experimental preview of the architecture behind Qwen4: a 125B-parameter MoE with 6B active per token, plus a 51B n-gram embedding table. This antirez DS4 Q2 build is a single 137.10 GiB GGUF with 41.73 GiB of resident main/MTP weights; its original 95.37 GiB BF16 n-gram table remains on a fast local SSD and is read row-by-row without being mapped or preloaded. It is the practical DS4 variant for Apple-silicon Macs, starting at a smaller context on memory-constrained systems.

Gezel downloads this model the first time a gezel needs it, checks the file against its published hash, and applies tuned settings for it.

The model’s own page · Every model · Model scorecard