# GLM 5.2

Frontier coding-first model with 1M context. 753B-parameter MoE with dual reasoning modes. MIT-licensed weights available.

## Overview

- **Provider:** ZAI
- **Model ID:** `glm-5.2`
- **Type:** Chat
- **Canonical URL:** https://modelstack.cc/models/chat/glm-5.2
- **Version:** 5.2
- **Best for:** Long-horizon coding, repo-scale agent work, complex engineering, agentic workflows with extended context

## Pricing

| Usage | Price |
| --- | ---: |
| Input | $1.4 / 1M tokens |
| Output | $4.4 / 1M tokens |
| Cache read | $0.26 / 1M tokens |
| Cache write (5 minutes) | $1.4 / 1M tokens |

## Specifications

- **Context window:** 1M
- **Maximum output:** 131K
- **Typical latency:** 18 ms
- **Released:** 2026-06-13

## Capabilities

- chat
- tool_use
- reasoning
- coding
- analysis

## API usage

Use this model ID in a ModelStack API request:

```json
{
  "model": "glm-5.2"
}
```

For the rendered page, visit https://modelstack.cc/models/chat/glm-5.2.