GLM 5.3 Flash

Model ID: vllm/release/glm-5-3-flash Z.AI (Zhipu) Experimental

EU AI Act

Attribute	Value
Classification	GPAI (General-Purpose AI Model)
Risk class	Limited Risk (Article 50)
Transparency obligation	Yes
noris role	Distributor

As a GPAI model, this model falls within the scope of the EU AI Act transparency obligations. noris acts as a distributor and makes the model developer’s information available pursuant to Article 53(2) and (3).

Technical Specifications

Property	Value
Model name	GLM 5.3 Flash
Provider	Z.AI (Zhipu)
Model ID	vllm/release/glm-5-3-flash
Parameters	320B (MoE), 18B active
Architecture	MoE with hybrid attention (sparse + linear), hyper connections (mHC)
Context length	1,048,576 tokens
Modalities	Text + image (multimodal)
Checkpoint	NVFP4, quantized by Red Hat AI (LLM Compressor)
License	MIT

Recommended Use Cases

• Coding and agentic engineering
• Long-context tasks, cost-efficient thanks to hybrid attention
• Image understanding
• Reasoning with adjustable effort (reasoning_effort: low, high, max)

Known Limitations

Caution

• Large model (high latency): due to its size, inference is slower than with smaller models.
• Primarily EN/ZH: performance is mainly optimized for English and Chinese.
• Experimental tier (2-week deprecation notice): this model may be deprecated with only two weeks advance notice.
• Controlling the thinking mode: reasoning_effort accepts low, high, and max and falls back to max when omitted. In chat scenarios, explicitly set clear_thinking=true.

Risk Assessment (EU AI Act)

Risk Factor	Category	Level
Model size	Deepfake potential	Medium
Open source (MIT)	Misuse potential	Low
Multimodal	Image input	Medium
Experimental tier	Deprecation notice	Flagged
Environmental impact	Model size (320B)	Medium-High

Training Data & Sustainability

Training data: publicly available in the technical report (arxiv: 2602.15763).

Sustainability:

Metric	Value
CO₂	Medium-High (320B MoE, 18B active)
Optimization	hybrid attention markedly lowers long-context serving costs

Original Sources

• Base model: https://huggingface.co/zai-org/GLM-5.3-Flash
• NVFP4 checkpoint: https://huggingface.co/RedHatAI/GLM-5.3-Flash-NVFP4
• License: MIT License
• Technical report: https://arxiv.org/abs/2602.15763

Model card created on October 7, 2026. This model card serves EU AI Act transparency purposes (Regulation (EU) 2024/1689). The information is based on publicly available data from the model provider.