GLM-4.6 & GLM-4.5

中文阅读

👋 Join our WeChat or Discord community.
📖 Check out the GLM-4.6 technical blog, technical report(GLM-4.5), and Zhipu AI technical documentation.
📍 Use GLM-4.6 API services on Z.ai API Platform.
👉 One click to GLM-4.6.

Model Introduction

GLM-4.6

Compared with GLM-4.5, GLM-4.6 brings several key improvements:

• Longer context window: The context window has been expanded from 128K to 200K tokens, enabling the model to handle more complex agentic tasks.
• Superior coding performance: The model achieves higher scores on code benchmarks and demonstrates better real-world performance in applications such as Claude Code、Cline、Roo Code and Kilo Code, including improvements in generating visually polished front-end pages.
• Advanced reasoning: GLM-4.6 shows a clear improvement in reasoning performance and supports tool use during inference, leading to stronger overall capability.
• More capable agents: GLM-4.6 exhibits stronger performance in tool using and search-based agents, and integrates more effectively within agent frameworks.
• Refined writing: Better aligns with human preferences in style and readability, and performs more naturally in role-playing scenarios.

We evaluated GLM-4.6 across eight public benchmarks covering agents, reasoning, and coding. Results show clear gains over GLM-4.5, with GLM-4.6 also holding competitive advantages over leading domestic and international models such as DeepSeek-V3.1-Terminus and Claude Sonnet 4.

GLM-4.5

The GLM-4.5 series models are foundation models designed for intelligent agents. GLM-4.5 has 355 billion total parameters with 32 billion active parameters, while GLM-4.5-Air adopts a more compact design with 106 billion total parameters and 12 billion active parameters. GLM-4.5 models unify reasoning, coding, and intelligent agent capabilities to meet the complex demands of intelligent agent applications.

Both GLM-4.5 and GLM-4.5-Air are hybrid reasoning models that provide two modes: thinking mode for complex reasoning and tool usage, and non-thinking mode for immediate responses.

We have open-sourced the base models, hybrid reasoning models, and FP8 versions of the hybrid reasoning models for both GLM-4.5 and GLM-4.5-Air. They are released under the MIT open-source license and can be used commercially and for secondary development.

As demonstrated in our comprehensive evaluation across 12 industry-standard benchmarks, GLM-4.5 achieves exceptional performance with a score of 63.2, in the 3rd place among all the proprietary and open-source models. Notably, GLM-4.5-Air delivers competitive results at 59.8 while maintaining superior efficiency.

For more eval results, show cases, and technical details, please visit our technical report or technical blog.

The model code, tool parser and reasoning parser can be found in the implementation of transformers, vLLM and SGLang.

Model Downloads

You can directly experience the model on Hugging Face or ModelScope or download the model by following the links below.

Model	Download Links	Model Size	Precision
GLM-4.6	🤗 Hugging Face 🤖 ModelScope	355B-A32B	BF16
GLM-4.5	🤗 Hugging Face 🤖 ModelScope	355B-A32B	BF16
GLM-4.5-Air	🤗 Hugging Face 🤖 ModelScope	106B-A12B	BF16
GLM-4.5-FP8	🤗 Hugging Face 🤖 ModelScope	355B-A32B	FP8
GLM-4.5-Air-FP8	🤗 Hugging Face 🤖 ModelScope	106B-A12B	FP8
GLM-4.5-Base	🤗 Hugging Face 🤖 ModelScope	355B-A32B	BF16
GLM-4.5-Air-Base	🤗 Hugging Face 🤖 ModelScope	106B-A12B	BF16

System Requirements

Inference

We provide minimum and recommended configurations for "full-featured" model inference. The data in the table below is based on the following conditions:

1. All models use MTP layers and specify --speculative-num-steps 3 --speculative-eagle-topk 1 --speculative-num-draft-tokens 4 to ensure competitive inference speed.
2. The cpu-offload parameter is not used.
3. Inference batch size does not exceed 8.
4. All are executed on devices that natively support FP8 inference, ensuring both weights and cache are in FP8 format.
5. Server memory must exceed 1T to ensure normal model loading and operation.

The models can run under the configurations in the table below:

Model	Precision	GPU Type and Count	Test Framework
GLM-4.5	BF16	H100 x 16 / H200 x 8	sglang
GLM-4.5	FP8	H100 x 8 / H200 x 4	sglang
GLM-4.5-Air	BF16	H100 x 4 / H200 x 2	sglang
GLM-4.5-Air	FP8	H100 x 2 / H200 x 1	sglang

Under the configurations in the table below, the models can utilize their full 128K context length:

Model	Precision	GPU Type and Count	Test Framework
GLM-4.5	BF16	H100 x 32 / H200 x 16	sglang
GLM-4.5	FP8	H100 x 16 / H200 x 8	sglang
GLM-4.5-Air	BF16	H100 x 8 / H200 x 4	sglang
GLM-4.5-Air	FP8	H100 x 4 / H200 x 2	sglang

if you are using AMD GPUs, Check here for AMD GPU deployment documentation.

Fine-tuning

The code can run under the configurations in the table below using Llama Factory:

Model	GPU Type and Count	Strategy	Batch Size (per GPU)
GLM-4.5	H100 x 16	Lora	1
GLM-4.5-Air	H100 x 4	Lora	1

The code can run under the configurations in the table below using Swift:

Model	GPU Type and Count	Strategy	Batch Size (per GPU)
GLM-4.5	H20 (96GiB) x 16	Lora	1
GLM-4.5-Air	H20 (96GiB) x 4	Lora	1
GLM-4.5	H20 (96GiB) x 128	SFT	1
GLM-4.5-Air	H20 (96GiB) x 32	SFT	1
GLM-4.5	H20 (96GiB) x 128	RL	1
GLM-4.5-Air	H20 (96GiB) x 32	RL	1

Quick Start

Both GLM-4.5 and GLM-4.6 use the same inference method.

First, install the required packages according to requirements.txt.

pip install -r requirements.txt

transformers

Please refer to the trans_infer_cli.py code in the inference folder.

vLLM

• Both BF16 and FP8 can be started with the following code:

vllm serve zai-org/GLM-4.5-Air \
    --tensor-parallel-size 8 \
    --tool-call-parser glm45 \
    --reasoning-parser glm45 \
    --enable-auto-tool-choice \
    --served-model-name glm-4.5-air

If you're using 8x H100 GPUs and encounter insufficient memory when running the GLM-4.5 / GLM-4.6 model, you'll need --cpu-offload-gb 16 (only applicable to vLLM).

If you encounter flash infer issues, use VLLM_ATTENTION_BACKEND=XFORMERS as a temporary replacement. You can also specify TORCH_CUDA_ARCH_LIST='9.0+PTX' to use flash infer (different GPUs have different TORCH_CUDA_ARCH_LIST values, please check accordingly).

SGLang

• BF16

python3 -m sglang.launch_server \
  --model-path zai-org/GLM-4.5-Air \
  --tp-size 8 \
  --tool-call-parser glm45  \
  --reasoning-parser glm45 \
  --speculative-algorithm EAGLE \
  --speculative-num-steps 3 \
  --speculative-eagle-topk 1 \
  --speculative-num-draft-tokens 4 \
  --mem-fraction-static 0.7 \
  --served-model-name glm-4.5-air \
  --host 0.0.0.0 \
  --port 8000

• FP8

python3 -m sglang.launch_server \
  --model-path zai-org/GLM-4.5-Air-FP8 \
  --tp-size 4 \
  --tool-call-parser glm45  \
  --reasoning-parser glm45  \
  --speculative-algorithm EAGLE \
  --speculative-num-steps 3  \
  --speculative-eagle-topk 1  \
  --speculative-num-draft-tokens 4 \
  --mem-fraction-static 0.7 \
  --disable-shared-experts-fusion \
  --served-model-name glm-4.5-air-fp8 \
  --host 0.0.0.0 \
  --port 8000

• PD-Disaggregation

The following is a simple method to implement PD-Disaggregation using a single machine with multiple GPUs, P and D each use 4 GPUs.

python -m sglang.launch_server --model-path zai-org/GLM-4.5-Air  --disaggregation-mode prefill --disaggregation-ib-device mlx5_0 --tp-size 4
python -m sglang.launch_server --model-path zai-org/GLM-4.5-Air  --disaggregation-mode decode --port 30001 --disaggregation-ib-device mlx5_0 --tp-size 4 --base-gpu-id 4                                                                                                                                                        
python -m sglang_router.launch_router --pd-disaggregation --prefill http://127.0.0.1:30000 --decode http://127.0.0.1:30001 --host 0.0.0.0 --port 8000

Request Parameter Instructions

• When using vLLM and SGLang, thinking mode is enabled by default when sending requests. If you want to disable the thinking switch, you need to add the extra_body={"chat_template_kwargs": {"enable_thinking": False}} parameter.
• Both support tool calling. Please use OpenAI-style tool description format for calls.
• For specific code, please refer to api_request.py in the inference folder.

Evaluation

• For tool-integrated reasoning, please refer to this doc.
• For search benchmark, we design a specific format for searching toolcall in thinking mode to support search agent, please refer to this. for the detailed template.