GLM-5

Best-in-class open-source model for systems engineering and long-horizon agents

Try Now

read docs

About model

GLM-5 is an open-source foundation model built for complex systems engineering and long-horizon agent workflows. It delivers production-grade productivity in large-scale programming tasks, with performance aligned with top closed-source models, and is designed for expert developers who think and build at the system level. Purpose-built for multi-stage, long-step complex tasks, GLM-5 autonomously decomposes system-level requirements with an architect-level approach while maintaining context coherence across automated workflows that run for hours.

Total Parameters (40B activated)

744B

MoE architecture for complex systems engineering

SWE-Bench Verified

77.8%

Best-in-class open-source coding

Vending Bench 2 (Open-Source)

Long-horizon planning and resource management

Model key capabilities

Agentic Long-Horizon Planning: Autonomously decomposes system-level requirements with architect-level approach, maintaining context coherence across workflows running for hours
Deep Debugging & Self-Correction: Analyzes logs, identifies root causes, and iteratively fixes compilation or runtime failures until the system runs end-to-end
Backend Architecture Excellence: Strong depth reasoning in backend architecture design, complex algorithm implementation, and difficult bug resolution
Opus-Level Intelligence: Benchmarks against Claude Opus 4.6 in code logic density and systems engineering capability with open-source flexibility and cost efficiency

Quickstart guides

RAG

Building a RAG Workflow

Agents

Agent Workflows

Apps

Next.js Chat Quickstart

Performance benchmarks

Model	GPQA Diamond	HLE	LiveCodeBench	MATH500	SWE-bench verified
GLM-5	86.0%				77.8%
Related open-source models
Competitor closed-source models
Claude Opus 4.6	90.5%	34.2%			78.7%
OpenAI o3	83.3%	24.9%		99.2%	62.3%
OpenAI o1	76.8%			96.4%	48.9%
GPT-4o	49.2%	2.7%	32.3%	89.3%	31.0%

API usage

cURL
Python
Typescript

Endpoint:

zai-org/GLM-5

curl -X POST "https://api.together.xyz/v1/chat/completions" \
  -H "Authorization: Bearer $TOGETHER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "zai-org/GLM-5",
    "messages": [
      {
        "role": "user",
        "content": "What are some fun things to do in New York?"
      }
    ]
}'

from together import Together

client = Together()

response = client.chat.completions.create(
  model="zai-org/GLM-5",
  messages=[
    {
      "role": "user",
      "content": "What are some fun things to do in New York?"
    }
  ]
)
print(response.choices[0].message.content)

import Together from 'together-ai';
const together = new Together();

const completion = await together.chat.completions.create({
  model: 'zai-org/GLM-5',
  messages: [
    {
      role: 'user',
      content: 'What are some fun things to do in New York?'
     }
  ],
});

console.log(completion.choices[0].message.content);

Model card
Architecture Overview:
• Open-source foundation model built for complex systems engineering and long-horizon agent workflows
• Designed for expert developers who think and build at the system level
• FP8 quantization for efficient inference
• Performance aligned with top closed-source models like Claude Opus 4.6
• Purpose-built for multi-stage, long-step complex tasks requiring deep reasoning

Training Methodology:
• Trained on large-scale programming tasks and system-level engineering workflows
• Emphasis on deep system construction and long-range agent execution
• Optimized for backend architecture design and complex algorithm implementation
• Self-reflection and error correction training for iterative debugging
• Architect-level decomposition of system requirements

Performance Characteristics:
• Agentic Long-Horizon Planning: Autonomously decomposes system-level requirements with architect-level approach
• Context Coherence: Maintains goal alignment across automated workflows running for hours
• Backend Refactoring: Strong depth reasoning in backend architecture design and complex algorithm implementation
• Deep Debugging: Analyzes logs, identifies root causes, iteratively fixes compilation/runtime failures
• Self-Correction: Robust error correction mechanisms ensuring end-to-end system execution
• Opus-Level Intelligence: Code logic density and systems engineering capability benchmarking against Claude Opus 4.6
• Open-Source Flexibility: Production deployment options with strong cost efficiency
‍
Applications & use cases
Large-Scale Systems Engineering:
• Complex backend architecture design and refactoring
• Multi-service system construction with deep integration requirements
• Large codebase modernization and migration projects
• Distributed systems design and implementation
• Microservices architecture planning and execution
• System-level performance optimization and scaling

Long-Horizon Agent Workflows:
• Automated workflows running for hours with maintained goal alignment
• Multi-stage deployment pipelines with autonomous error recovery
• Complex CI/CD orchestration with self-correction mechanisms
• End-to-end system automation requiring deep reasoning
• Infrastructure-as-code generation and management

Deep Debugging & Error Resolution:
• Root cause analysis of complex compilation failures
• Runtime error diagnosis across distributed systems
• Log analysis and issue identification in large-scale applications
• Iterative debugging with self-reflection mechanisms
• Production incident resolution requiring system-level understanding

Backend Architecture & Algorithms:
• Complex algorithm implementation and optimization
• Database schema design and query optimization
• API design and backend service architecture
• Performance-critical code generation and refactoring
• System bottleneck identification and resolution

Expert Developer Workflows:
• Architect-level system decomposition and planning
• Technical debt reduction in enterprise codebases
• Legacy system modernization with maintained functionality
• Code review and architectural consultation
• System design documentation and technical specifications

Open-Source Alternative Deployment:
• Teams requiring Opus-level intelligence with open-source flexibility
• Cost-efficient production deployment for systems engineering tasks
• On-premises or private cloud deployment with full model control
• Custom fine-tuning for domain-specific systems engineering
• Research and development requiring model transparency
‍

Related models

Model specifications

Model data

Model provider
ZAI
Type
Chat
Reasoning
LLM
Main use cases
Chat
Function Calling
Features
Function Calling
JSON Mode
Speed
High
Intelligence
Very High
Deployment
Serverless
Monthly Reserved
Endpoint
zai-org/GLM-5
Parameters
744B
Context length
202K
Input price
$1.00 / 1M tokens
Output price
$3.20 / 1M tokens
Input modalities
Text
Output modalities
Text

Released
February 10, 2026
Last updated
February 12, 2026
Quantization level
FP4
External link
Provider docs
Category
Chat

Run in Playground

Quickstart docs

Deploy model

GLM-5

About model

API usage

Model card

Applications & use cases