---
title: GLM-5.2-SF-INT8
canonical_url: "https://www.modelscope.cn/models/siliconflow/GLM-5.2-SF-INT8"
md_url: "https://www.modelscope.cn/models/siliconflow/GLM-5.2-SF-INT8.md"
repository: siliconflow/GLM-5.2-SF-INT8
chinese_name: siliconflow/GLM-5.2-SF-INT8
last_updated: 2026-07-22
license: mit
pipeline_tag: text-generation
tasks:
  - text-generation
model_type:
  - glm_moe_dsa
architectures:
  - GlmMoeDsaForCausalLM
parameters: 756.8B
tensor_type:
  - I8
  - BF16
  - F32
library_name:
  - safetensors
  - pytorch
frameworks:
  - pytorch
language:
  - en
  - zh
downloads: 361
stars: 0
---

# GLM-5.2-SF-INT8

> GLM-5.2-SF-INT8 - siliconflow 在 ModelScope 开源的模型。GLM-5.2-SF-INT8 is an INT8 checkpoint based on GLM-5.2, published by SiliconFlow for efficient text generation and deployment.

siliconflow/GLM-5.2-SF-INT8 是 ModelScope 魔搭社区上的 756.8B 参数text-generation模型，采用 mit 许可。

- **Repository**: siliconflow/GLM-5.2-SF-INT8
- **License**: mit
- **Tasks**: text-generation
- **Parameters**: 756.8B
- **Downloads**: 361
- **Stars**: 0
- **Last updated**: 2026-07-22

Source: https://www.modelscope.cn/models/siliconflow/GLM-5.2-SF-INT8

---

# GLM-5.2-SF-INT8

GLM-5.2-SF-INT8 is an INT8 checkpoint based on **GLM-5.2**, published by SiliconFlow for efficient text generation and deployment.

## Model information

- **Base model:** GLM-5.2
- **Task:** Text generation
- **Languages:** Chinese and English
- **Weight format:** Safetensors
- **License:** MIT

The tokenizer, chat template, and generation configuration are included in this repository.

## Usage

Use an inference runtime that supports the GLM-5.2 architecture and this INT8 checkpoint format. Please follow the runtime's model-loading and serving documentation.

For chat applications, use the included tokenizer and chat template.

## Notes

Model behavior, capabilities, and limitations generally follow the GLM-5.2 base model. Users should evaluate the model for their target workload before production deployment.