GLM 4.5 Air
GLM 4.5 Air by Z AI (Zhipu AI) is a Mixture-of-Experts model with 128K context, dual reasoning modes for thinking and direct responses, and MIT licensing.
Model Overview
Capabilities, design details, and architectural traits
GLM 4.5 Air - Efficient Agentic MoE Model
GLM 4.5 Air is Z.AI's (Zhipu AI) compact Mixture-of-Experts model for intelligent-agent workloads. The official model card positions it as a lighter member of the GLM-4.5 family that keeps the same agentic design goals.
Defining Characteristics
- Agent-oriented design: The model is documented as part of the GLM-4.5 series for intelligent-agent applications.
- Dual response behavior: It supports both thinking mode for complex reasoning and tool use, and non-thinking mode for immediate responses.
- Compact alternative: The model card frames GLM 4.5 Air as the more compact alternative to the full GLM-4.5 model.
- Commercial-ready release: The MIT license allows commercial use and secondary development.
Positioning
GLM 4.5 Air exists for workloads that benefit from GLM-4.5 family capabilities but need a smaller model footprint. Officially, it is the efficiency tier in the series rather than a separate product line.
Benchmark Performance
Independent evaluations · Artificial Analysis
Accuracy & Capability Details
Compare Models Side-by-Side
Evaluate specifications, pricing, and independent benchmark indices
| Model Details | |||
|---|---|---|---|
| General Info | |||
| Provider | Z AI | Anthropic | Anthropic |
| Release Date | July 28, 2025 | September 22, 2026 | September 28, 2026 |
| Knowledge Cutoff | Dec 2024 | - | Jun 2026 |
| Context & Limits | |||
| Context Window | 131K | 1M Best Context Window | 1M Best Context Window |
| Pricing (per 1M tokens) | |||
| Input Pricing | $0.17 Best Input Pricing | $4 | $2 |
| Output Pricing | $0.98 Best Output Pricing | $20 | $10 |
| Modalities | |||
| Inputs | text | textimagefile | textimagefile |
| Outputs | text | text | text |
| Benchmarks (0-100) | |||
| Intelligence Index | 11.1 | 57.6 Best Intelligence Index | 56.0 |
| Coding Index | - | - | - |
| Agentic Index | - | - | - |
Humanity's Last Exam
Extremely difficult logical reasoning and knowledge.
Long Context Reasoning
Logical reasoning over long context windows.
SciCode Benchmark
Scientific coding and mathematical modeling.
Independent evaluation data provided by Artificial Analysis. To view the latest benchmarks and full details, visit their official site.
Explore more from Z AI
Other models by Z AI
Top AI Models
Leading alternatives by intelligence score
