Model•active

GLM-5.3-Flash

GLM-5.3-Flash is Z.ai's 320B-parameter open-weight efficiency model with 18B active parameters, a 1,048,576-token context window, and a hybrid sparse/linear-attention multimodal architecture.

Official siteVerified Oct 5, 2026

Overview

Flash starts from a newly trained base model and combines sparse attention, linear attention, mixture-of-experts routing, and multimodal pre-training. The official processor configuration includes image and video processing.

Model specifications

Family: GLM-5.3
Context: 1,048,576 tokens
Input: Text, Image, Video
Output: Text
Open weights: Yes
Capabilities: Reasoning, Vision

Related entities

Belongs To

Model Familyactive

GLM-5.3

GLM-5.3 is Z.ai's current open-weight GLM generation for coding, reasoning, long-horizon agentic tasks, and efficient multimodal inference.

Related by Develops

Companyactive

Z.ai

Z.ai develops the GLM model family and publishes open-weight coding, reasoning, multimodal, and agent-oriented models.