GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
XReyRobert
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
XReyRobert
Model Tree
Jackrong/Qwopus3.8-27B-Flash
Input Modalities
Output Modalities
Supported Functionality
desc_act=false, true sequential0.25, beta 0.22.00.05, auto increment 0.01The calibration artifact is identified by SHA-256
5d8d9a2fe6ecc47dc937358784b5698564d3b217deaf13dab418c7c2fb06eaaa.
The upload workflow fails closed unless all of the following pass:
The source tensors are preserved so the artifact is complete, but preservation alone does not establish MTP speculative-decoding or vision runtime compatibility. Treat those paths as experimental until separately validated with the target serving stack.
The source model card reports faster decoding and high MTP acceptance for its own GGUF/llama.cpp evaluation. Those figures are not claimed for this GPTQ-Pro artifact. Publish measured GPTQ/vLLM results only after a controlled run on the target hardware.