What are the hardware requirements for GLM 4.6 FP8?

GLM 4.6 FP8 requires approximately 166.3 GB of VRAM and supports a context window of 198K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed GLM 4.6 FP8 and how large is it?

GLM 4.6 FP8 is developed by zai-org, a model with 358.5b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

Where can I download or evaluate GLM 4.6 FP8?

GLM 4.6 FP8 is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

GLM 4.6 FP8 by zai-org — VRAM 166.3GB, 198K context

Name: GLM 4.6 FP8
Author: zai-org

GLM 4.6 FP8 is an open-source language model by zai-org. Features: 358.5b LLM, VRAM: 166.3GB, Context: 198K, License: mit, LLM Explorer Score: 0.27.

Arxiv:2508.06471 Compressed-tensors Conversational Deploy:azure En Endpoints compatible Eval-results Glm4 moe Region:us Safetensors Sharded Tensorflow Zh

Model Card on HF 🤗: https://huggingface.co/zai-org/GLM-4.6-FP8

GLM 4.6 FP8 Parameters and Internals

LLM Name	GLM 4.6 FP8
Repository 🤗	https://huggingface.co/zai-org/GLM-4.6-FP8
Model Size	358.5b
Required VRAM	166.3 GB
Updated	2026-08-08
Maintainer	zai-org
Model Type	glm4_moe
Model Files	1.9 GB: 1-of-92 0.3 GB: 2-of-92 0.3 GB: 3-of-92 3.9 GB: 4-of-92 3.9 GB: 5-of-92 3.9 GB: 6-of-92 3.9 GB: 7-of-92 3.9 GB: 8-of-92 3.9 GB: 9-of-92 3.9 GB: 10-of-92 3.9 GB: 11-of-92 3.9 GB: 12-of-92 3.9 GB: 13-of-92 3.9 GB: 14-of-92 3.9 GB: 15-of-92 3.9 GB: 16-of-92 3.9 GB: 17-of-92 3.9 GB: 18-of-92 3.9 GB: 19-of-92 3.9 GB: 20-of-92 3.9 GB: 21-of-92 3.9 GB: 22-of-92 3.9 GB: 23-of-92 3.9 GB: 24-of-92 3.9 GB: 25-of-92 3.9 GB: 26-of-92 3.9 GB: 27-of-92 3.9 GB: 28-of-92 3.9 GB: 29-of-92 3.9 GB: 30-of-92 3.9 GB: 31-of-92 3.9 GB: 32-of-92 3.9 GB: 33-of-92 3.9 GB: 34-of-92 3.9 GB: 35-of-92 3.9 GB: 36-of-92 3.9 GB: 37-of-92 3.9 GB: 38-of-92 3.9 GB: 39-of-92 3.9 GB: 40-of-92 3.9 GB: 41-of-92 3.9 GB: 42-of-92 3.9 GB: 43-of-92 3.9 GB: 44-of-92 3.9 GB: 45-of-92
Supported Languages	en zh
Model Architecture	Glm4MoeForCausalLM
License	mit
Context Length	202752
Model Max Length	202752
Transformers Version	4.54.0
Vocabulary Size	151552
Torch Data Type	bfloat16

Rank the GLM 4.6 FP8 capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

Best Alternatives to GLM 4.6 FP8

Best Alternatives	Context / RAM	Downloads	Likes
GLM 4.7 FP8	198K / 166.3 GB	37391	124
GLM 4.7 FP8	198K / 166.3 GB	14	3
GLM 4.5 FP8	128K / 167.8 GB	891	78

Note: green Score (e.g. "73.2") means that the model is better than zai-org/GLM-4.6-FP8.