GLM-4.5
About GLM-4.5
The GLM-4.5 series models are foundation models designed for intelligent agents, by unifying reasoning, coding, and intelligent agent capabilities to meet the complex demands of intelligent agent applications. GLM-4.5 has 355 billion total parameters with 32 billion active parameters, providing two modes: thinking and non-thinking.
Unlock the full potential of intelligent agents with GLM-4.5. Its unified reasoning, coding, and agentic capabilities empower solutions for complex, real-world challenges, leveraging both deep 'thinking' and rapid 'non-thinking' modes.
Autonomous Agent Orchestration
Design and deploy self-improving agents that plan, execute, and adapt to complex tasks using GLM-4.5's unified reasoning and coding.
Use Case Example:
"An agent autonomously managed a cloud infrastructure, identifying performance bottlenecks, writing Python scripts to scale resources, and deploying fixes without human intervention."
Agent-Driven Code Refactoring
Enable agents to analyze codebases, understand architectural patterns, and perform complex refactoring or generate new modules with high-level reasoning.
Use Case Example:
"An agent refactored a legacy Java enterprise application into a modern Spring Boot microservice, optimizing for modularity and scalability by reasoning about data flow and dependencies."
Strategic Business Intelligence
Develop agents that synthesize diverse business data (reports, market trends, customer feedback) to provide multi-faceted strategic recommendations and predict outcomes.
Use Case Example:
"An agent analyzed quarterly financial statements, social media sentiment, and competitor news to generate a detailed market entry strategy for a new product, including risk assessments and projected ROI."
Automated System Design & Verification
Utilize GLM-4.5 to design robust system architectures, verify logical consistency across components, and generate configuration scripts for deployment.
Use Case Example:
"An agent designed a resilient Kubernetes deployment for a high-traffic e-commerce platform, generating Helm charts and validating network policies for security and scalability."
Dynamic Knowledge Synthesis
Create agents that dynamically synthesize information from vast knowledge bases, explain complex concepts, and adapt learning paths based on user interaction and reasoning.
Use Case Example:
"An agent developed a personalized learning module for advanced machine learning concepts, drawing from multiple research papers and adapting explanations based on the user's prior knowledge and questions."
Metadata
Specification
State
Deprecated
Architecture
Calibrated
No
Mixture of Experts
Yes
Total Parameters
335B
Activated Parameters
32B
Reasoning
No
Precision
FP8
Context length
131K
Max Tokens
Compare with Other Models
See how this model stacks up against others.
Z.ai
chat
GLM-4.7
Release on: Dec 23, 2025
Total Context:
205K
Max output:
205K
Input:
$
0.42
/ M Tokens
Output:
$
2.2
/ M Tokens
Z.ai
chat
GLM-4.6V
Release on: Dec 8, 2025
Total Context:
131K
Max output:
131K
Input:
$
0.3
/ M Tokens
Output:
$
0.9
/ M Tokens
Z.ai
chat
GLM-4.6
Release on: Oct 4, 2025
Total Context:
205K
Max output:
205K
Input:
$
0.39
/ M Tokens
Output:
$
1.9
/ M Tokens
Z.ai
chat
GLM-4.5-Air
Release on: Jul 28, 2025
Total Context:
131K
Max output:
131K
Input:
$
0.14
/ M Tokens
Output:
$
0.86
/ M Tokens
Z.ai
chat
GLM-4.5V
Release on: Aug 13, 2025
Total Context:
66K
Max output:
66K
Input:
$
0.14
/ M Tokens
Output:
$
0.86
/ M Tokens
Z.ai
chat
GLM-4.1V-9B-Thinking
Release on: Jul 4, 2025
Total Context:
66K
Max output:
66K
Input:
$
0.035
/ M Tokens
Output:
$
0.14
/ M Tokens
Z.ai
chat
GLM-Z1-32B-0414
Release on: Apr 18, 2025
Total Context:
131K
Max output:
131K
Input:
$
0.14
/ M Tokens
Output:
$
0.57
/ M Tokens
Z.ai
chat
GLM-4-32B-0414
Release on: Apr 18, 2025
Total Context:
33K
Max output:
33K
Input:
$
0.27
/ M Tokens
Output:
$
0.27
/ M Tokens
Z.ai
chat
GLM-Z1-9B-0414
Release on: Apr 18, 2025
Total Context:
131K
Max output:
131K
Input:
$
0.086
/ M Tokens
Output:
$
0.086
/ M Tokens
