GLM-5.2 vs GLM-5.3-Flash

Best cell per row highlighted. Null means undisclosed — never counted as zero.

Z.ai
GLM-5.2GA
8.8
AI Panel
Z.ai
GLM-5.3-FlashGA
8.6
AI Panel

Identity & lifecycle

ProviderZ.aiZ.ai
Family / tierGLM-5 · FlagshipGLM-5 · Flash
StatusGAGA
Released2026-06-162026-08-26
Knowledge cutoff——

Architecture & context

Context window1M1M
Max output tokens131K131K
Input modalitiestexttext, image, video
Reasoning modeoptionalalways
Open weightsYesYes

Pricing (per Mtok)

Input$1.4$0.15
Output$4.4$0.5
Cached input$0.26$0.03
Batch input——
Free tierNoNo

Speed

Output speed (tok/s)166.6—
Time to first token (s)1.39—
Latency tierfastmedium

Trust & deployment

Trains on inputs—No
LicenseMITMIT
Clouds——
Compliance——

AI Panel scoring

Unified score8.88.6
Decision Maker8.58.5
Domain Strategist99
Finance Lead9.59.5
Domain Practitioner99
Power User8.58.5
Skeptic87
Value score9.69.8

Benchmarks

GPQA Diamond91.2—
Terminal-Bench8184.3
Humanity's Last Exam40.5—
LMArena Elo14721475
LMArena Coding Elo15101525

Scores link to their sources. Missing cells mean the vendor hasn't disclosed a result — honesty over padding.