Zhipu AI
GLM-5.2
Open flagship GLM for long-horizon coding agents and million-token context work
zhipuai/glm-5.2
Context
1M
Output
131.1K
Released
2026-06-13
Knowledge cutoff
—
Types
Reasoning
Tool use
Structured output
Attachments
Open weights
Provider pricing
No providers currently list this model.
Benchmarks
SWE-Bench Pro
62.1 · resolve rate
Terminal-Bench
82.7 · success rate
FrontierSWE
74.4 · dominance
Humanity's Last Exam
40.5 · accuracy
Humanity's Last Exam
54.7 · accuracy
CritPt
20.9 · accuracy
AIME
99.2 · accuracy
HMMT
94.4 · accuracy
HMMT
92.5 · accuracy
IMOAnswerBench
91 · accuracy
GPQA Diamond
91.2 · accuracy
NL2Repo
48.9 · resolve rate
DeepSWE
46.2 · resolve rate
Program Bench
63.7 · score
Terminal-Bench
81 · success rate
PostTrainBench
34.3 · score
SWE Marathon
13 · resolve rate
MCP Atlas
76.8 · score
Tool-Decathlon
48.2 · score