Skip to main content
ZA

GLM-5.3 Flash

zhipuai/glm-5.3-flash

Z.AI multimodal GLM model with a hybrid sparse and linear attention architecture, 320B total and 18B activated parameters. It supports image, video, text and file input, text output, and long-context coding and professional workflows.

Compare
Context window
1M
Input
Document · Image · Text · Video
Output
Text
Released
2026-08-26
Max output
128K
Model parameters
320B
Active parameters
18B
Capabilities
Attachments, Reasoning, Tool call

Pricing

IN

$0.15

USD / 1,000,000 token

Cache read

$0.03

USD / 1,000,000 token

OUT

$0.50

USD / 1,000,000 token

Performance

Overall capability

53.5/100

Coding
52.1
Reasoning
46.5
Professional work
56.3
Knowledge & language
51.5
Finance AgentProfessional workVals AI2026-11-05
Humanity’s Last ExamGLM-5.3-FlashReasoningArtificial Analysis2026-11-05
AA-Analyst AgentGLM-5.3-FlashProfessional workArtificial Analysis2026-11-05
AA-BriefcaseProfessional workArtificial Analysis2026-11-05
GDPval-AAProfessional workArtificial Analysis2026-11-05
ProgramBenchCodingVals AI2026-11-05
Terminal-Bench Science · AAGLM-5.3-FlashReasoningArtificial Analysis2026-11-05
APEX-Agents 1.1Professional workMercor2026-11-05

Providers

No providers found

ModelRegistry · Catalog updated 2026-11-05