GLM-4.5-Air
Active Parameters
106B
Context Length
128K
Modality
Multimodal
Architecture
Mixture of Experts
KV Cache Rate
184 KB/tok
License
MIT License
Release Date
28 Jul 2025
Knowledge Cutoff
Mar 2025
API Pricing (per 1M)
Input: $0.13 · Output: $0.85
Hardware Sizing
Precision
Context
Workload Memory: 223.73 GB
FP16 · 8K context
Base Weights: 219.79 GB
KV Cache: 1.44 GB
Runtime & Graphs: 2.50 GB
Minimum Required Hardware
8x NVIDIA H100 80GB
223.7 / 576.0 GB usable · TP 8
Determine requirements for your specific workload:
Open in VRAM Calculator
Architecture Diagram
Evaluation Benchmarks
Rank
#118
| Benchmark | Score | Rank |
|---|---|---|
Professional Knowledge | 0.814 | 26 |
Graduate-Level QA | 0.75 | 76 |
General Text | 1373 | 128 |
Intelligence Index | 0.11 | 231 |
Rankings
Overall Rank
#118
Coding Rank
-
About GLM-4.5-Air
GLM-4.5-Air is a high-efficiency 106B Mixture-of-Experts model by Z.ai activating 12B parameters for responsive agentic workflows and code generation. It incorporates Multi-Token Prediction and hybrid reasoning modes across a 128K context window.
Technical Specifications
Attention
Attention Structure
Multi-Head Attention
Attention Heads
96
Key-Value Heads
8
Attention Head Dimension
128
Position Embedding
Absolute Position Embedding
RoPE Theta
1,000,000
Normalization
RMS Normalization
Activation Function
Swish
Dimensions
Hidden Dimension Size
4,096
Number of Layers
46
FFN Intermediate Size (Dense)
1,408
Multi-Token Prediction Heads
1
Tokenizer
Vocabulary Size
151,552
Mixture of Experts
Total Expert Parameters
12.0B
Number of Experts
129
Active Experts
9
Shared Experts
1
FFN Intermediate Size (per Expert)
1,408
Dense Layers Before MoE
1
About GLM Family
General Language Models from Z.ai