Atlas

Compare Models

Claude Haiku 5.5 vs Muse Wonder Full

Muse Wonder Full leads by 3.3 capability points (≈95% interval +1.0 to +5.7): a clear lead over Claude Haiku 5.5 · P(higher) ≈ 100%

Wins 70 of 79 shared benchmarks (7 lost, 2 tied)

Capability vs. relative cost
Claude Haiku 5.5Muse Wonder FullClosed weightsOpen weights

Each line runs through one model’s priced reasoning settings, cheapest to most expensive; grey points are other models’ priced settings in the same range. Results whose reasoning effort the source didn’t report are left out wherever a model has named effort levels. Capability and relative cost are the published estimates (relative cost is cost per task as a multiple of GPT-5 (high)); hover a setting, or use the arrow keys, to set it beside the other model’s same setting.

Claude Haiku 5.5vsMuse Wonder Fullcapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins7 – 70 · 79 shared, 2 tied

Comparing matching setups only (1 benchmark excluded) · suites counted once
← Claude Haiku 5.5Muse Wonder Full →

Claude Haiku 5.5vsMuse Wonder Fullwins by categoryData Nov 15, 2026Data: Vals AI, Artificial Analysis, Arena and 6 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Muse Wonder Full minus Claude Haiku 5.5.

LisanBench Difficulty-WeightedGames27119337+71.0
LisanBench Path LengthGames1289732190+59.9
Vending-Bench 2Games360212292+52.6
AutomationBench-AA 1.0.6Agents36.071.3+35.4
IOI (Vals v2)Code47.379.8+32.6
Terminal-Bench Science 0.1 (Vals)Science10.037.1+27.1
Artificial Analysis Coding Agent Index v1.5Indexes41.468.0+26.6
Terminal-Bench 4.0Code36.458.1+21.7
Harvey's Legal Agent BenchmarkLegal1.321.7+20.4
Bug Hunt BenchCode20.540.0+19.5
HLAB-Vals-1.1Legal4.223.3+19.2
Humanity's Last Exam DiamondGeneral QA29.648.5+18.9
Artificial Analysis Strategy & Ops IndexIndexes39.157.9+18.8
Terminal-Bench-Science 0.1Science20.037.1+17.1
LMArena Code Arena WebDevCode15941723+17.0
Code MigrationCode44.661.0+16.5
LMArena Agent Arena Net Improvement (Steering Burden)Agents0.59.0+16.0
Artificial Analysis Healthcare & Medical IndexIndexes37.551.9+14.4
Artificial Analysis Intelligence Index v5.0.1Indexes36.550.1+13.6
Artificial Analysis Intelligence Index v5.0Indexes36.249.7+13.5
Excel Modeling BenchmarkFinance & Accounting59.071.9+12.9
Tax Agent BenchFinance & Accounting62.275.0+12.8
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & Medicine35.022.2−12.8
Artificial Analysis Engineering IndexIndexes42.555.3+12.7
Artificial Analysis Intelligence Index v4.3.2Indexes43.455.8+12.4
Finance Agent v2 - General Qualitative AnalysisFinance & Accounting61.974.2+12.3
Artificial Analysis Legal IndexIndexes38.750.7+12.1
GDP.pdf (Artificial Analysis, All-pass)Professional Work20.832.6+11.8
Artificial Analysis Finance & Accounting IndexIndexes43.655.2+11.7
Artificial Analysis Economics IndexIndexes42.754.1+11.4
CritPt v2 (agentic, avg@5)Science64.075.1+11.1
Humanity's Last Exam (Text-Only)General QA42.753.6+10.9
Public Benefits Bench v1.1Professional Work63.273.7+10.6
Code Migration v1.1Code52.462.8+10.4
CritPt (avg@5)Science22.632.9+10.3
Vals Index v2.1Indexes54.464.6+10.1
GDPval-AA v2.1Professional Work16291857+10.1
Vals Science IndexIndexes38.548.5+10.0
AA-Omniscience IndexIndexes10.730.1+9.7
Legal Research BenchLegal43.853.4+9.6
Excel Modeling Benchmark v1.1Finance & Accounting65.575.0+9.5
LMArena Agent Arena Praise vs ComplaintAgents-3.515.4+9.5
Immigration BenchLegal61.871.2+9.4
AA-Briefcase v1.1 EloAgents15731729+8.6
Vals Index v2.2Indexes58.766.8+8.1
Finance Agent v2 - PrecedentsFinance & Accounting27.035.0+8.1
SciCode 1.0.1 (Artificial Analysis)Code55.062.6+7.6
Vals Legal IndexIndexes70.377.9+7.6
Finance Agent v2 - Disclosure AnalysisFinance & Accounting64.371.6+7.3
Finance Agent v2 - AdjustmentsFinance & Accounting49.656.7+7.1
Vals Finance IndexIndexes57.464.6+7.1
Finance Agent v2 - All PassFinance & Accounting42.949.0+6.2
SRE Bench — Fully SolvedCode5.711.8+6.1
Artificial Analysis Progress Index v1Indexes326343+6.0
Finance Agent v2 - General Quantitative AnalysisFinance & Accounting74.480.1+5.7
AA-LCR v1.1Search78.383.7+5.3
Finance Agent v2 - Earnings AnalysisFinance & Accounting73.578.6+5.1
ProofBench v1.1Math89.094.0+5.0
LMArena Agent Arena Confirmed SuccessAgents1.611.1+4.8
Finance Agent v2 - Financial ModelingFinance & Accounting28.323.7−4.6
Vibe Code Bench v1.1Code90.486.0−4.4
GDPxProfessional Work4.99.0+4.1
LMArena Agent Arena Steering BurdenAgents2.49.6+3.6
SWE-bench Multimodal (Vals v1, 480 Tasks)Code44.047.3+3.3
SWE-Marathon v1.1Code50.053.1+3.1
FrontierCode 1.1 Main (Artificial Analysis)Code41.844.9+3.1
BioMysteryBench (Vals Terminus 2)Science70.473.3+3.0
FrontierCode 2.0 (Artificial Analysis)Code34.937.8+2.8
Finance Agent v2.1Finance & Accounting65.467.6+2.2
ProgramBenchCode2.50.5−2.0
CyberBench v1.1 OverallCybersecurity75.573.6−1.8
Finance Agent v2 - Market AnalysisFinance & Accounting63.164.9+1.8
MedCodeHealth & Medicine51.549.9−1.6
Finance Agent v2 - ComparablesFinance & Accounting45.346.8+1.5
LMArena Agent Arena Tool HallucinationAgents-2.10.0+1.0
LMArena Agent Arena Bash ReliabilityAgents7.48.3+0.4
MedScribeHealth & Medicine89.589.3−0.2
MysteryMechanismScience35.135.10.0
VulcanBench Frontier v4 (functional pass@1)Code95.795.70.0

Claude Haiku 5.5vsMuse Wonder Fullnormalised score marginsData Nov 15, 2026Data: Vals AI, Artificial Analysis, Arena and 6 othersVillage LLM Atlasatlas.village.dev

Win counts use each benchmark’s declared score direction. Represented suite slices and the derived Capability score are excluded from headline totals. Missing or incompatible setups are excluded, never counted as losses.
Loading Atlas data…