GPT-6.1 Sol

编程65.88 项

• ProgramBenchVals AI3.0%折算分 至多 83.3+
• Terminal-Bench 4.0Vals AI55.1%折算分 75.2+
• Code MigrationVals AI65.1%折算分 73.7+
• Terminal-Bench 4.0Artificial Analysis56.1%折算分 72.4+
• Vibe Code Bench v1.1Vals AI88.9%折算分 70.3+
• LiveBench CodingLiveBench80.4%折算分 60.1+
• SciCodeArtificial Analysis54.2%折算分 54.6+
• CyberBench Patch v1.1Vals AI78.6%折算分 52.5+

科研74.79 项

• Mystery Game PuzzlesEpoch AI80.0%折算分 85.0+
• FrontierMath T1–3Epoch AI93.7%折算分 83.2+
• Terminal-Bench ScienceVals AI52.9%折算分 82.7+
• Terminal-Bench ScienceArtificial Analysis58.1%折算分 81.6+
• Chess PuzzlesEpoch AI68.0%折算分 77.6+
• MysteryMechanismVals AI46.4%折算分 75.5+
• CritPtArtificial Analysis31.7%折算分 71.8+
• LiveBench ReasoningLiveBench92.6%折算分 70.8+
• Humanity's Last ExamArtificial Analysis52.9%折算分 67.9+

专业办公65.310 项

• LiveBench Data AnalysisLiveBench82.7%折算分 71.8+
• AA-Analyst AgentArtificial Analysis50.0%折算分 70.0+
• AA-BriefcaseArtificial Analysis1,555折算分 69.7+
• AutomationBench-AAArtificial Analysis64.9%折算分 68.3+
• APEX-AgentsMercor60.0%折算分 67.9+
• τ³-BankingMercor43.0%折算分 65.4+
• GDPval-AAArtificial Analysis1,596.2折算分 65.1+
• MedCodeVals AI48.8%折算分 60.1+
• Finance AgentVals AI52.0%折算分 59.1+
• Harvey 法律 Agent 基准Vals AI5.4%折算分 56.5+

知识问答74.27 项

• MLCR-AAArtificial Analysis38.3%折算分 91.5+
• SimpleQA VerifiedEpoch AI73.9%折算分 88.0+
• LiveBench LanguageLiveBench90.1%折算分 84.7+
• AA-OmniscienceArtificial Analysis41.5折算分 81.1+
• LiveBench IFLiveBench74.2%折算分 69.2+
• PBullshitBench v2BullshitBench59.0%折算分 68.5+
• AA-LCRArtificial Analysis83.0%折算分 61.7+

真人盲选66.61 项

• Arena TextLMArena1,479.9折算分 61.8+

尚无成绩的评测

这些评测尚无可采用的公开成绩，缺测不补分。