GLM-5.2

GLM-5.2 is positioned as a large-scale reasoning model from Z.ai, with openly released weights hosted on Hugging Face, which lets teams self-host and fine-tune for proprietary workflows. It is described as particularly strong at coding and tool use, with the ability to hold engineering context together through an entire development workflow, moving from requirements to multi-platform deployment inside a single task. Two reasoning effort levels are supported, with the higher setting mapping to maximum reasoning depth, which makes the model adaptable to both quick interactive completions and careful, step-by-step problem solving.

In practical terms, GLM-5.2 is aimed at project-level software engineering and complex multi-step automation rather than short, single-turn exchanges. Its strengths in maintaining context and following standards consistently across long-running tasks make it a natural fit for agent frameworks, automated code generation pipelines, and orchestrated tool-calling systems where reliability over many steps matters more than raw single-shot speed. The combination of open weights and agent-oriented design gives teams flexibility to deploy it on their own infrastructure while leveraging it for sustained, goal-driven work.

ClinePasscline-pass/glm-5.2glm

Open provider page Provider docs

Quick Info

Provider: ClinePass
Model key: cline-pass/glm-5.2
Release date: Jun 13, 2026
Last updated: Jun 13, 2026
Input modalities
Output modalities
Capabilities

Cost

Input token cost: $1.40
Output token cost: $4.40

Limits

Output tokens: 131,072 tokens
Context window: 1,000,000 tokens

Transparent token rates

Compare GLM-5.2 pricing

Rates are shown per one million tokens. Combined means one million input plus one million output tokens.

Browse this family

Latest news about GLM-5.2

ClinePass

CoverageAnalysis

A third-party technical analysis reports the launch of GLM-5.2, a next-generation open-source coding model positioned as an iterative upgrade of the GLM-5 series. The article frames the release as a focused effort on complex engineering workflows, long-context agent tasks, and computational efficiency, with reasoning p Architecturally, the article describes GLM-5.2 as retaining an ultra-large-parameter design from the GLM-5 series — 744 billion total parameters with around 40 billion activated per inference via dynamic parameter activation — and highlights an integrated DeepSeek Sparse Attention (DSA) mechanism as the core breakthrou

ClinePass

Coverage

Simon Willison's June 17, 2026 review provides independent context for GLM-5.2, noting that Z.ai released it to coding-plan subscribers on June 13 and then open-weighted the model under an MIT license on June 16. He describes it as a 753B-parameter, 1.51TB Mixture-of-Experts model with 40 active parameters, text-input On pricing and efficiency, Willison cites OpenRouter providers charging roughly $1.40 per million input and $4.40 per million output tokens for GLM-5.2, far below GPT-5.5 and Claude Opus 4.5-4.8, while flagging that Artificial Analysis found the model unusually token-hungry at 43k output tokens per Intelligence Index t