GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
local-inference-lab
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
local-inference-lab
Model Tree
Qwen/Qwen3.8-27B
Input Modalities
Output Modalities
Supported Functionality
| Component | Representation |
|---|---|
| MLP gate, up and down projections | NVFP4, 16-element blocks |
| Gated delta network projections | MXFP8, 32-element blocks; frozen |
| Full-attention query, key, value and output projections | Original BF16; frozen |
| Student LM head | Trained FP32 master rounded to BF16 |
| Text normalization weights | Trained FP32 masters |
| Token embeddings, GDN convolutions and dynamics | Original BF16; frozen |
| Vision encoder and remaining source tensors | Unchanged |
Packed NVFP4 and MXFP8 weights reconstruct to the BF16 weight values used during training. The tokenizer, chat template, generation configuration and multimodal processors are retained from the base model.
The 192 MLP input scales are copied exactly from Qwen3.8-27B-QAD-E1,
whose step-4,779 weights were calibrated on 390,497,191 raw-text and chat
tokens. That calibration selected the pooled p99.999 token-row maximum
from exact BF16 histograms. These scales were not recalibrated at step 7,000.
Each dense layer has equal gate/up scales in separate tensors and an independent down-projection scale. Serving uses calibrated NVFP4 MLP activations and dynamic MXFP8 GDN activations, adding activation quantization beyond the BF16 training forward.
Hugging Face safetensors with ModelOpt mixed-precision metadata. The runtime must support the base architecture, NVFP4 dense linears, MXFP8 linears and the BF16 full-attention/head exclusions. Packed reconstruction, serialized tensors and copied input scales are checked. Serving quality and performance have not been evaluated for this export.
Apache 2.0, following the base model.