What are the hardware requirements for GLM 4.7 Flash?

GLM 4.7 Flash requires approximately 57.3 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 4.7 Flash and how large is it?

GLM 4.7 Flash is developed by zai-org, a model with 31.2b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does GLM 4.7 Flash perform on standard benchmarks?

GLM 4.7 Flash has the following published scores: LMSYS ELO 1368. Compare against reference models on this page or on the LLM Explorer leaderboards.

GLM 4.7 Flash by zai-org — VRAM 57.3GB, 198K context

Name: GLM 4.7 Flash
Author: zai-org

GLM 4.7 Flash is an open-source language model by zai-org. Features: 31.2b LLM, VRAM: 57.3GB, Context: 198K, License: mit, MoE, LLM Explorer Score: 0.57, ELO: 1368.

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

Model Card on HF 🤗: https://huggingface.co/zai-org/GLM-4.7-Flash

GLM 4.7 Flash Benchmarks

LMSys ELO 1368 vs 1272 (so35) 7.5%

^nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").

What is the LLM Explorer Rank (Score)

GLM 4.7 Flash Parameters and Internals

LLM Name	GLM 4.7 Flash
Repository 🤗	https://huggingface.co/zai-org/GLM-4.7-Flash
Model Size	31.2b
Required VRAM	57.3 GB
Updated	2026-07-28
Maintainer	zai-org
Model Type	glm4_moe_lite
Model Files	1.4 GB: 1-of-48 1.3 GB: 2-of-48 1.3 GB: 3-of-48 1.3 GB: 4-of-48 1.3 GB: 5-of-48 1.3 GB: 6-of-48 1.3 GB: 7-of-48 1.3 GB: 8-of-48 1.3 GB: 9-of-48 1.3 GB: 10-of-48 1.3 GB: 11-of-48 1.3 GB: 12-of-48 1.3 GB: 13-of-48 1.3 GB: 14-of-48 1.3 GB: 15-of-48 1.3 GB: 16-of-48 1.3 GB: 17-of-48 1.3 GB: 18-of-48 1.3 GB: 19-of-48 1.3 GB: 20-of-48 1.3 GB: 21-of-48 1.3 GB: 22-of-48 1.3 GB: 23-of-48 1.3 GB: 24-of-48 1.3 GB: 25-of-48 1.3 GB: 26-of-48 1.3 GB: 27-of-48 1.3 GB: 28-of-48 1.3 GB: 29-of-48 1.3 GB: 30-of-48 1.3 GB: 31-of-48 1.3 GB: 32-of-48 1.3 GB: 33-of-48 1.3 GB: 34-of-48 1.3 GB: 35-of-48 1.3 GB: 36-of-48 1.3 GB: 37-of-48 1.3 GB: 38-of-48 1.3 GB: 39-of-48 1.3 GB: 40-of-48 1.3 GB: 41-of-48 1.3 GB: 42-of-48 1.3 GB: 43-of-48 1.3 GB: 44-of-48
Supported Languages	en zh
Model Architecture	Glm4MoeLiteForCausalLM
License	mit
Context Length	202752
Model Max Length	202752
Transformers Version	5.0.0rc0
Vocabulary Size	154880

Rank the GLM 4.7 Flash 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

Quantized Models of the GLM 4.7 Flash

Model	Likes	Downloads	VRAM
GLM 4.7 Flash AWQ 4bit	56	489953	20 GB
GLM 4.7 Flash Abliteratex	3	510	62 GB
GLM 4.7 Flash 8bit	24	3470	31 GB
GLM 4.7 Flash 4bit	66	7029	17 GB
GLM 4.7 Flash 6bit	7	4053	24 GB
GLM 4.7 Flash 4bit Mxfp4	4	237	15 GB
GLM 4.7 Flash 8bit Gs32	5	183	33 GB

Best Alternatives to GLM 4.7 Flash

Best Alternatives	Context / RAM	Downloads	Likes
GLM 4.7 Flash	198K / 58.6 GB	295368	15
Mochi	198K / 58.6 GB	75	2
...ihui GLM 4.7 Flash Abliterated	198K / 58.6 GB	1196	84
GLM 4.7 Flash Derestricted	198K / 58.6 GB	156	34