Compare / GPT-5.4 Nano vs GLM 5.1

GPT-5.4 Nano vs GLM 5.1

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 Nano
OpenAI · profile
GLM 5.1
Z.ai · profile
Results compared
20
Shared sources
3
BenchmarkMetricGPT-5.4 NanoGLM 5.1
Arena (formerly LMArena)reportedOverallArena rating 1,4181,456
Arena (formerly LMArena)reportedBusiness, management and financeArena rating 1,4051,465
Arena (formerly LMArena)reportedCreative writingArena rating 1,3351,448
Arena (formerly LMArena)reportedExpert promptsArena rating 1,4381,495
Arena (formerly LMArena)reportedInstruction followingArena rating 1,3871,460
Arena (formerly LMArena)reportedOverallArena rating 1,4011,465
Arena (formerly LMArena)reportedWriting, literature and languageArena rating 1,3621,453
Artificial AnalysisreportedArtificial Analysis Coding Indexindex score 56.155.8
Artificial AnalysisreportedArtificial Analysis Intelligence Indexindex score 20.726.1
Artificial AnalysisreportedGPQA Diamond% of questions 81.7%86.8%
Artificial AnalysisreportedHumanity's Last Exam% of questions 28.3%30.1%
Artificial AnalysisreportedIFBench% of instructions 75.9%76.3%
Artificial AnalysisreportedLong-context reasoning (AA-LCR)% of questions 76.7%73.7%
Artificial AnalysisreportedSciCode% of problems 47.2%44.8%
Artificial AnalysisreportedTerminal-Bench 2.1% of tasks 60.7%61.8%
Artificial AnalysisreportedTerminal-Bench 4.0% of tasks 0.5%2.0%
Artificial AnalysisreportedTerminal-Bench Hard% of tasks 42.4%43.2%
Artificial AnalysisreportedΤ-bench banking% of tasks 27.4%13.6%
Artificial AnalysisreportedΤ²-bench telecom% of tasks 76.0%97.7%
SimpleQA Verified (Epoch AI)reportedCorrect answers% of questions 11.7%34.0%

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.