Downloads · 30 days
10
16% of all-time downloads
GLM-5.2-HYBRID-FP8-MXFP4
long10024070/GLM-5.2-HYBRID-FP8-MXFP4
GLM-5.2-HYBRID-FP8-MXFP4 is a text generation model from long10024070. Use it when you need the model to write or continue text. It is set up for transformers. The card lists the license as mit.
This deterministic checkpoint uses block-FP8 non-MoE weights and the block-FP8 MTP layer 78 from zai-org/GLM-5.2-FP8. Routed and shared MoE expert tensors for layers 3 through 77 are copied byte-for-byte from amd/GLM-…
All-time downloads
62
Public
Parameters
390B
414 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors414 GB · 100%
Parameter types
How the weights are stored.
U8364B · 93%
Try a prompt
Base models
Model card
From the Hugging Face model README
GLM-5.2 Hybrid FP8/MXFP4
This deterministic checkpoint uses block-FP8 non-MoE weights and the block-FP8
MTP layer 78 from zai-org/GLM-5.2-FP8. Routed and shared MoE expert tensors
for layers 3 through 77 are copied byte-for-byte from
amd/GLM-5.2-MXFP4. No tensor is dequantized or requantized during assembly.
The checkpoint declares hybrid-fp8-mxfp4 and requires the matching conditional
SGLang loader patch shipped with its reproducibility assets. Original FP8 and
Quark checkpoints do not use that patch path.
- Task
- Text Generation
- Library
- transformers
- Type
- glm_moe_dsa
- License
- mit
- Created
- Aug 27, 2026
- Updated
- Aug 27, 2026