glm-5.3#

  • Context Length: 1048576

  • Model Name: glm-5.3

  • Languages: en, zh

  • Abilities: chat, tools, reasoning

  • Description: GLM-5.3 is a reasoning model for coding and long-horizon agentic tasks.

Specifications#

Model Spec 1 (pytorch, 753 Billion)#

  • Model Format: pytorch

  • Model Size (in billions): 753

  • Quantizations: none

  • Engines: vLLM, Transformers

  • Model ID: zai-org/GLM-5.3-BF16

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name glm-5.3 --size-in-billions 753 --model-format pytorch --quantization ${quantization}

Model Spec 2 (fp8, 753 Billion)#

  • Model Format: fp8

  • Model Size (in billions): 753

  • Quantizations: FP8

  • Engines: vLLM

  • Model ID: zai-org/GLM-5.3

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name glm-5.3 --size-in-billions 753 --model-format fp8 --quantization ${quantization}