GLM-5 is Z.ai's (formerly Zhipu AI) flagship open-source foundation model, built on a 744-billion-parameter Mixture-of-Experts architecture with 40 billion active parameters. It is engineered for complex systems design and long-horizon agent workflows, delivering production-grade performance on large-scale programming tasks that rivals leading closed-source models. With advanced agentic planning, deep backend reasoning, and iterative self-correction, GLM-5 moves beyond code generation to full-system construction and autonomous execution.
API|Reasoning|Open ModelMIT
Knowledge Cutoff
2026-02-17
Input → Output Format
Context Memory
80KIN131KOUT
AI Performance Evaluation
Arena Overall Score
1456
±5As of 2026-05-01
Overall Rank
No.32
19,980 Votes
Arena by Ability
Hard Prompts
1477±6No.37
Expert Knowledge
1483±15No.32
Instruction Following
1445±8No.36
Conversation Memory
1466±11No.36
Creative
1444±11No.21
Coding
1493±9No.46
Math
1446±16No.39
Arena by Occupation
Creative Writing
1442±9No.33
Social Sciences
1482±10No.27
Media
1432±10No.31
Business
1448±10No.41
Healthcare
1488±16No.20
Legal
1465±15No.34
Software
1486±7No.44
Mathematics
1462±18No.32
Source:Arena Intelligence
Overall
AA Intelligence Index
50%↑11%
LiveBench
69%↑8%
ForecastBench
60%↑1%
Reasoning & Math
GPQA Diamond
82%↑0%
HLE
27%↑10%
LB Reasoning
69%↑0%
LB Math
83%↑9%
LB Data
68%↑15%
Coding
AA Coding Index
44%↑8%
LB Coding
74%↑1%
LB Agentic
55%↑10%
TAU2
98%↑18%
TerminalBench
43%↑9%
SciCode
46%↑4%
Language & Instructions
IFBench
72%↑9%
AA-LCR
63%↑1%
Hallucination (HHEM)
10%↑0%
Factual (HHEM)
90%↑0%
LB Language
78%↑5%
LB IF
55%↑4%
Output Speed
Standard Mode
50tok/s↓27
First Output 1.36s
Reasoning Mode
60tok/s↓27
First Output 60.17s
