Z.ai·GLM

GLM-5.3

Kept the GLM-5.2 base and improved complex coding, long-horizon tasks, and cyber capability entirely through scaled post-training.

CurrentReasoningCodingAgentWeights not released

MODEL NUMBER028

Specifications

Only published specifications are shown.

Architecture

MoE

Input modalities

Text

Open weights

Status

Current

Kept the GLM-5.2 base and improved complex coding, long-horizon tasks, and cyber capability entirely through scaled post-training.

GLM-5.3 is a clear case of major agent gains coming from scaled post-training without changing the base model.

Compared with GLM-5.2 ↗

PreviousGLM-5.2— · 1M ctx

CurrentGLM-5.3— · — ctx

New capabilities Adjustable reasoning Cybersecurity

Adjustable reasoningCodingCybersecurityLong-horizon agent

2026-08-14Announcement

GLM-5.3 launched with gains attributed entirely to scaled post-training.

Select a card to view the related model.