Z.ai
Z.ai

GLM-5

Model ID
2026-02-11Open Model

GLM-5 is Z.ai’s flagship open-source foundation model engineered for complex systems design and long-horizon agent workflows. Built for expert developers, it delivers production-grade performance on large-scale programming tasks, rivaling 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
Knowledge Cutoff
2026-02-17
Input → Output Format
Context Memory
80KIN131KOUT
Cost/1M Words
$0.72IN$2.3OUT

AI Performance Evaluation

Arena Overall Score
1456
±6
As of 2026-04-07
Overall Rank
No.23
13,401 Votes
Arena by Ability
Hard Prompts
1476±7No.27
Expert Knowledge
1484±19No.24
Instruction Following
1448±10No.24
Conversation Memory
1461±13No.30
Creative
1445±13No.17
Coding
1489±11No.38
Math
1458±20No.20
Arena by Occupation
Creative Writing
1441±11No.24
Social Sciences
1487±13No.10
Media
1431±12No.25
Business
1448±11No.30
Healthcare
1499±20No.9
Legal
1468±19No.21
Software
1482±9No.34
Mathematics
1469±22No.18
Overall
AA Intelligence Index
50%↑11%
LiveBench
69%↑11%
ForecastBench
60%↑0%
Reasoning & Math
GPQA Diamond
82%↑1%
HLE
27%↑11%
LB Reasoning
69%↑10%
LB Math
83%↑12%
LB Data
68%↑18%
Coding
AA Coding Index
44%↑10%
LB Coding
74%↑0%
LB Agentic
55%↑13%
TAU2
98%↑27%
TerminalBench
43%↑12%
SciCode
46%↑5%
Language & Instructions
IFBench
72%↑17%
AA-LCR
63%↑2%
LB Language
78%↑7%
LB IF
55%↑4%
Output Speed
Standard Mode
53tok/s↓26
First Output 1.33s
Reasoning Mode
69tok/s↓62
First Output 50.96s