Skip to content

wrldsuksgo2mars

Qwen3.8-Flash-Next-EXL3-K4.25-v1

wrldsuksgo2mars/Qwen3.8-Flash-Next-EXL3-K4.25-v1

Qwen3.8-Flash-Next-EXL3-K4.25-v1 is a machine learning model from wrldsuksgo2mars. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as other.

Calibrated mixed K4/K5 EXL3 routed experts, averaging 4.25 bits per expert weight, including all 48 target layers and the native MTP layer. Per layer, K5 upgrades cover 72 gate, 120 up and 192 down projections; remain…

Downloads · 30 days

94

24% of all-time downloads

All-time downloads

400

Public

Parameters

89.7B

180 GB on disk

Likes

2

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors179 GB · 100%

Parameter types

How the weights are stored.

BF1656.7B · 63%

Base models

Type
qwen4_exp
License
other
Created
Sep 7, 2026
Updated
Sep 7, 2026