Compare / GPT-5.4 Mini vs GLM 5.2

GPT-5.4 Mini vs GLM 5.2

20 results from 3 sources that measured both models. Each row is in the source's own units; there is no overall winner.

GPT-5.4 Mini
OpenAI · profile
GLM 5.2
Z.ai · profile
Results compared
20
Shared sources
3
BenchmarkMetricGPT-5.4 MiniGLM 5.2
Arena (formerly LMArena)reportedOverallArena rating 1,4471,463
Arena (formerly LMArena)reportedBusiness, management and financeArena rating 1,4601,463
Arena (formerly LMArena)reportedCreative writingArena rating 1,4021,455
Arena (formerly LMArena)reportedExpert promptsArena rating 1,4801,500
Arena (formerly LMArena)reportedInstruction followingArena rating 1,4341,470
Arena (formerly LMArena)reportedOverallArena rating 1,4471,476
Arena (formerly LMArena)reportedWriting, literature and languageArena rating 1,4241,459
Artificial AnalysisreportedArtificial Analysis Coding Indexindex score 56.168.8
Artificial AnalysisreportedArtificial Analysis Intelligence Indexindex score 24.133.7
Artificial AnalysisreportedGPQA Diamond% of questions 87.5%89.5%
Artificial AnalysisreportedHumanity's Last Exam% of questions 28.1%41.1%
Artificial AnalysisreportedIFBench% of instructions 73.3%73.3%
Artificial AnalysisreportedLong-context reasoning (AA-LCR)% of questions 77.0%78.3%
Artificial AnalysisreportedSciCode% of problems 52.1%51.2%
Artificial AnalysisreportedTerminal-Bench 2.1% of tasks 59.2%77.9%
Artificial AnalysisreportedTerminal-Bench 4.0% of tasks 2.0%1.0%
Artificial AnalysisreportedTerminal-Bench Hard% of tasks 52.3%50.8%
Artificial AnalysisreportedΤ-bench banking% of tasks 25.6%34.6%
Artificial AnalysisreportedΤ²-bench telecom% of tasks 83.3%99.1%
SimpleQA Verified (Epoch AI)reportedCorrect answers% of questions 29.4%34.2%

Bold green marks the better value on that metric. Where a source reports ranges that overlap, the difference may not be meaningful; see the benchmark page for ranges.