Muse Spark 1.2

编程58.210 项

• ProgramBenchVals AI0.5%折算分 至多 80.3+
• Vibe Code Bench v1.1Vals AI79.1%折算分 63.5+
• CyberBench Patch v1.1Vals AI85.7%折算分 62.7+
• LiveBench Agentic CodingLiveBench57.6%折算分 61.4+
• SciCodeArtificial Analysis57.4%折算分 59.4+
• APEX-SWEMercor40.5%折算分 57.6+
• LiveBench CodingLiveBench77.5%折算分 57.5+
• Terminal-Bench 4.0Artificial Analysis7.1%折算分 52.0+
• PFrontierSWE v2Proximal Labs14.3%折算分 49.6+
• Terminal-Bench 4.0Vals AI6.1%折算分 46.7+

科研61.05 项

• LiveBench ReasoningLiveBench90.0%折算分 65.9+
• CritPtArtificial Analysis17.7%折算分 63.0+
• Humanity's Last ExamArtificial Analysis45.5%折算分 62.4+
• BioMysteryBenchVals AI64.8%折算分 61.8+
• Chess PuzzlesEpoch AI27.0%折算分 52.7+

专业办公61.88 项

• TaxAgentBenchVals AI32.8%折算分 67.0+
• Legal Research BenchVals AI43.8%折算分 66.9+
• LiveBench Data AnalysisLiveBench76.5%折算分 62.9+
• GDPval-AAArtificial Analysis1,480折算分 61.9+
• AA-BriefcaseArtificial Analysis1,324.8折算分 61.3+
• MedCodeVals AI49.3%折算分 61.0+
• AutomationBench-AAArtificial Analysis40.6%折算分 58.0+
• APEX-AgentsMercor45.4%折算分 57.2+

知识问答60.77 项

• SimpleQA VerifiedEpoch AI60.3%折算分 68.6+
• LiveBench IFLiveBench74.3%折算分 66.8+
• LNYT Connections 扩展版Lech Mazur79.2%折算分 59.4+
• LiveBench LanguageLiveBench78.6%折算分 58.6+
• PBullshitBench v2BullshitBench49.0%折算分 57.9+
• AA-LCRArtificial Analysis79.0%折算分 52.7+
• MLCR-AAArtificial Analysis7.2%折算分 38.3+

真人盲选60.82 项

• Arena TextLMArena1,492.2折算分 66.5+
• Arena WebDevLMArena1,533.7折算分 55.9+

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。