Atlas

Compare Models

Kimi K3.1 Pro vs Muse Wonder Full

Muse Wonder Full leads by 0.8 capability points (≈95% interval −1.6 to +3.2): statistically tied · P(higher) ≈ 75%

Wins 33 of 40 shared benchmarks (7 lost)

Capability vs. relative cost
Kimi K3.1 ProMuse 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.

Kimi K3.1 ProvsMuse Wonder Fullcapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins7 – 33 · 40 shared

Comparing matching setups only (42 benchmarks excluded) · suites counted once
← Kimi K3.1 ProMuse Wonder Full →

Kimi K3.1 ProvsMuse Wonder Fullwins by categoryData Nov 15, 2026Data: Vals AI, Arena, Artificial Analysis and 4 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 Kimi K3.1 Pro.

Legal Research Bench Business/CommercialLegal27.958.1+30.2
Legal Research Bench CriminalLegal28.149.1+21.1
Legal Research Bench Weighted PassLegal71.489.8+18.4
Legal Research Bench FamilyLegal13.631.8+18.2
Legal Research Bench ImmigrationLegal18.236.4+18.2
Legal Research Bench Constitutional/Civil RightsLegal31.647.4+15.8
SRE Bench — Fully SolvedCode27.111.8−15.3
Legal Research Bench Civil LitigationLegal29.644.4+14.8
HLAB-Vals-1.1Legal11.723.3+11.7
Tax Agent BenchFinance & Accounting63.975.0+11.1
Humanity's Last Exam DiamondGeneral QA39.448.5+9.1
MysteryMechanismScience26.135.1+9.0
Vals Finance IndexIndexes56.064.6+8.5
LMArena Agent Arena Net Improvement (Steering Burden)Agents4.99.0+7.7
Legal Research Bench Administrative/RegulatoryLegal56.163.6+7.6
Legal Research Bench HealthLegal73.366.7−6.7
Finance Agent v2.1Finance & Accounting61.567.6+6.1
Artificial Analysis Coding Agent Index v2.1Indexes48.654.5+5.8
SWE-Marathon - release-reported version unspecifiedCode51.757.5+5.8
Bug Hunt BenchCode34.340.0+5.7
Terminal-Bench Science 0.1 (Vals)Science31.437.1+5.7
Terminal-Bench 4.0Code63.658.1−5.6
Access to Justice BenchmarkLegal69.574.7+5.2
Excel Modeling Benchmark v1.1Finance & Accounting70.075.0+5.0
LMArena Code Arena WebDevCode16881723+4.7
LMArena Agent Arena Confirmed SuccessAgents2.711.1+4.2
Vals Science IndexIndexes44.448.5+4.2
LMArena Agent Arena Praise vs ComplaintAgents8.015.4+3.7
ProofBench v1.1Math91.094.0+3.0
AA-Briefcase v1.1 EloAgents16781729+2.8
RiemannBench (November 2026 rescoring)Math60.863.2+2.4
BioMysteryBench (Vals Terminus 2)Science75.673.3−2.2
IOI (Vals v2)Code81.979.8−2.1
Code Migration v1.1 (Vals Index v2.2 subset)Code59.761.7+2.0
AutomationBench-AA 1.0.6Agents73.371.3−2.0
LMArena Agent Arena Steering BurdenAgents6.79.6+1.4
Vibe Code Bench v1.1Code84.986.0+1.1
LMArena Agent Arena Bash ReliabilityAgents7.28.3+0.5
GDPval-AA v2.1Professional Work18461857+0.5
LMArena Agent Arena Tool HallucinationAgents1.00.0−0.5

Kimi K3.1 ProvsMuse Wonder Fullnormalised score marginsData Nov 15, 2026Data: Vals AI, Arena, Artificial Analysis and 4 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…