Back to Models
Zhipu AIOpen Source
GLM4-MoE-100B-A10B
GLM4-MoE-100B-A10B is a foundation model developed by Zhipu AI. It is a chat-specialized model equipped with a large number of parameters and a long context window of 128K.
Parameters
1000.0B
Context Window
128K
License
Apache 2.0
Release Date
2025-07-14
API Pricing
API pricing for this model is not yet available
Strengths
- ・High expressiveness from massive parameters
- ・Supports long 128K context
- ・Open-source Apache 2.0 license
Weaknesses
- ・Extremely large 200GB model size
- ・Requires high-end computational resources
- ・Complex memory management with MoE architecture
Use Cases
- ・Developing advanced conversational AI chatbots
- ・Long document analysis and summarization
- ・Open-source-based AI research
Deep Analysis
Architecture
MoE (100B total, 10B active)
Earlier GLM-4 generation
Context Window
128K tokens
Release Date
2024
GLM-4-0414 generation
License
Open weights
Strengths
- ・Efficient MoE architecture
- ・10B active parameters for fast inference
- ・128K context
- ・Open weights
Weaknesses
- ・Older generation, superseded by GLM-4.5
- ・Less capable than newer GLM models
- ・Limited documentation in English
Competitor Comparison
| Model | Arena | SWE | GPQA | Price |
|---|---|---|---|---|
| GLM-4.5-Air 106B | - | - | - | Comparable |
| Qwen 2.5 72B | - | - | - | Comparable |
GLM4-MoE-100B-A10B is an earlier-generation MoE model from Zhipu AI with 100B total and 10B active parameters. It has been superseded by the GLM-4.5 family which offers significantly improved performance.
Analysis generated: 2026-05-24