GLM 5

Visit Tool

GLM 5 is a Coding & Development tool that provides a next-generation large language model with 745 billion parameters. It features a 44 billion active Mixture of Experts (MoE) architecture, accessible via Z.ai's platform and APIs.

Claim this tool

2Views

At a glance

Pricing

Usage-based · Open Source

Free tier

API

Yes

Skill level

Technical

About

What is GLM 5?

GLM-5 is Zhipu AI's fifth-generation large language model, boasting approximately 745 billion parameters within a Mixture of Experts (MoE) architecture, with 44 billion active parameters. Engineered for agentic intelligence, advanced multi-step reasoning, coding, and creative writing, it represents a significant leap in AI capability. The model is accessible via Z.ai's platform and WaveSpeed API, with an anticipated open-weight release under the MIT license. Notably, GLM-5 was trained entirely on Huawei Ascend chips using the MindSpore framework, demonstrating independence from US-manufactured hardware and positioning it as a direct challenger to models like OpenAI's GPT-5 and Anthropic's Claude.

Best used for

Ideal for developers who need to integrate frontier-level AI for advanced reasoning, generate high-quality code, and build autonomous agentic systems. Especially valuable for those seeking cost-effective, high-performance models with a focus on open-weight access and hardware independence.

Common actions

generate code

reason logically

create content

develop agents

process long contexts

Capabilities

Key features

745B total parameters
44B active MoE
200K context window
DeepSeek Sparse Attention
Huawei Ascend trained
Agentic intelligence
Advanced reasoning

Target Audience

developer

Integrations

Not yet documented

Pricing & Plans

Usage-based · Open Source

Not publicly disclosed. Check glm5.net for current pricing.

FAQs

What is the 'Pony Alpha' model and its connection to GLM-5?

In early February 2026, a mysterious model named 'Pony Alpha' appeared on OpenRouter. Speculation, fueled by GitHub pull requests and benchmark results, suggests it may be a stealth release or early version of GLM-5, though Zhipu AI has not officially confirmed this connection.

Will GLM-5 be open source and what is its pricing model?

GLM-5 is anticipated to be released as an open-weight model under the MIT license, enabling free commercial use and fine-tuning. API access is available via Z.ai's platform and WaveSpeed API at ultra-low costs, significantly more economical than competitors.

What hardware was used to train GLM-5 and why is it significant?

GLM-5 was trained entirely on Huawei Ascend chips using the MindSpore framework. This achieves full independence from US-manufactured semiconductor hardware, marking a significant milestone for China's domestic AI infrastructure and demonstrating its viability at frontier scale.

Trending

Subcategories trending in Coding & Development

Code Assistants DevOps & Infrastructure No-Code / Low-Code Testing & QA Backend & APIs Prompt Engineering

Trending

Also listed in

This tool also appears in

AI Agents & Automation › AI Frameworks & Infra AI Agents & Automation › General-Purpose Agents

Explore

Browse AI tools by category

Content & Design Productivity & Business Coding & Development AI Agents & Automation Research & Education Wellness & Lifestyle Career Development Marketing & Growth Data & Analytics Customer Support & CX Finance E-commerce