Atlas

Compare Models

Kimi K3.1 Pro vs Muse Wonder

Kimi K3.1 Pro leads by 0.3 capability points (≈95% interval −2.0 to +2.6): statistically tied · P(higher) ≈ 59%

Wins 51 of 92 shared benchmarks (41 lost)

Capability vs. relative cost
Kimi K3.1 ProMuse WonderClosed 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 Wondercapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins51 – 41 · 92 shared

Comparing matching setups only (59 benchmarks excluded) · suites counted once
← Kimi K3.1 ProMuse Wonder →

Kimi K3.1 ProvsMuse Wonderwins by categoryData Nov 15, 2026Data: Vals AI, Arena, Surge AI and 33 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Muse Wonder minus Kimi K3.1 Pro.

Visa Agentic Commerce BenchmarkAgents53.382.2+28.9
Legal Research Bench ImmigrationLegal18.245.5+27.3
Legal Research Bench FamilyLegal13.636.4+22.7
DAYJOB: Finance (October 2026 rescoring)Finance & Accounting17.038.8+21.8
Legal Research Bench Business/CommercialLegal27.948.8+20.9
CWE-bench v2 Pass@1Cybersecurity45.366.1+20.8
DAYJOB: AccountingFinance & Accounting13.630.8+17.2
Legal Research Bench Weighted PassLegal71.487.7+16.3
Legal Research Bench CriminalLegal28.143.9+15.8
IOI (Vals v2)Code81.966.9−14.9
Terminal-Bench 4.0Code63.649.5−14.1
DAYJOB: LawLegal23.936.6+12.7
HLAB-Vals-1.1Legal11.723.8+12.1
DAYJOB: Data ScienceProfessional Work4.716.7+12.0
FrontierML ($200 wallet)Agents29.619.3−10.3
micro1 PharmaBenchHealth & Medicine27.517.2−10.3
Code Migration v1.1 (Vals Index v2.2 subset)Code59.769.8+10.1
DeepFinanceFinance & Accounting34.644.2+9.6
HoneyBench v0.2 (hack rate)Safety71.881.0−9.2
Harvey LAB (Mercor run, undisclosed ~213-task cohort)Legal13.823.0+9.2
Mystery Game PuzzlesGames42.051.0+9.0
Terminal-Bench Science 0.1 (Vals)Science31.422.9−8.6
BioMysteryBench (Vals Terminus 2)Science75.667.0−8.5
Creative Writing v3 EloChat & Writing21732007−8.4
τ³-Banking (Mercor run, bm25_grep retrieval)Agents59.851.9−7.9
Legal Research Bench Constitutional/Civil RightsLegal31.639.5+7.9
Legal Research Bench Administrative/RegulatoryLegal56.163.6+7.6
MineBench — Bradley-Terry RatingGames19921868−7.5
Legal Research Bench Civil LitigationLegal29.637.0+7.4
Tax Agent BenchFinance & Accounting63.971.2+7.3
ProofBench v1.1Math91.084.0−7.0
FrontierCode 2.0 - Pass RateCode22.028.8+6.8
GameBench 2Games84.077.8−6.2
LMArena Agent Arena Net Improvement (Steering Burden)Agents4.91.8−5.9
FrontierCode 2.0 - ScoreCode19.925.3+5.4
Vals Science IndexIndexes44.439.1−5.2
Vals Finance IndexIndexes56.061.2+5.2
AutomationBench-AA 1.0.6Agents73.368.1−5.2
LMArena Code Arena WebDevCode16881649−5.1
Agents' Last Exam v1.1 (ALE v1.1)Agents43.438.5−4.9
LMArena Agent Arena Praise vs ComplaintAgents8.0-1.8−4.9
LMArena Text Arena EloChat & Writing14831511+4.9
FrontierMath Tier 4 (v2) - PrivateMath65.970.7+4.9
NoCoT-Bench NCKI (kspine v3)General QA116109−4.8
APEX-Accounting 1.1Finance & Accounting62.467.1+4.7
EnterpriseBench: CoreCraftProfessional Work66.161.5−4.6
FrontierMemory v1.1Agents55.651.1−4.4
Artificial Analysis Coding Agent Index v2.1Indexes48.644.5−4.2
Bug Hunt BenchCode34.330.2−4.1
Finance Agent v2.1Finance & Accounting61.565.5+4.1
FrontierMath Tiers 1-3 (v2) - PrivateMath78.282.1+3.9
Humanity's Last ExamGeneral QA51.147.6−3.5
Access to Justice BenchmarkLegal69.573.0+3.5
BuseyBench SVGMultimodal6.96.6−3.4
GDPval-AA v2.1Professional Work18461770−3.3
Legal Research Bench HealthLegal73.376.7+3.3
Real-SWE 1.1Code39.242.4+3.2
DualEntry Accounting AI BenchmarkFinance & Accounting86.183.2−2.9
NoCoT-Bench NCRI 15.2General QA10397.7−2.8
Excel Modeling Benchmark v1.1Finance & Accounting70.072.5+2.5
DAXBench v2Code86.284.0−2.1
LMArena Agent Arena Steering BurdenAgents6.72.6−2.1
Creative Writing v3 Rubric ScoreChat & Writing16.616.2−2.0
ErdosBenchMath3.03.0+2.0
MazeBench ASCIIGames2.00.0−2.0
DTBenchGeneral QA93.395.2+1.9
Vibe Code Bench v1.1Code84.983.1−1.8
Furniture Assembly BenchmarkMultimodal55.854.2−1.7
RiemannBench (November 2026 rescoring)Math60.859.2−1.6
SWE-Marathon - release-reported version unspecifiedCode51.753.1+1.4
LiveBench 2.0Agents76.778.1+1.4
MysteryMechanismScience26.127.5+1.4
ACCoRDGeneral QA77.576.2−1.3
LMArena Agent Arena Confirmed SuccessAgents2.75.4+1.3
LMArena Agent Arena Bash ReliabilityAgents7.24.8−1.2
APEX-Agents 1.1Professional Work75.977.0+1.1
AA-Briefcase v1.1 EloAgents16781659−1.1
FrontierBioMLScience25.826.9+1.0
RWE-benchAgents1.00.0−1.0
VulnPR-100Code47.046.0−1.0
OpenWorldEngineer (Mean@8)Code13.612.7−0.8
MMMU ProMultimodal84.184.9+0.8
LMCAGeneral QA57.656.8−0.8
LMArena Agent Arena Tool HallucinationAgents1.0-0.5−0.8
Prophet Arena Arena Score (Agentic, OpenCode)Forecasting86.185.4−0.7
sudo L7 (October 2026 rescoring)Code16.716.1−0.6
OTIS Mock AIME 2024-2025Math100.099.4−0.6
Humanity's Last Exam DiamondGeneral QA39.439.0−0.4
Prophet Arena Arena Score (Fixed Context)Forecasting85.485.2−0.2
Conceptual Reasoning IndexIndexes67.867.7−0.1
GBENCH One-shot Coding GScoreGames58.258.1−0.1
Epoch Capabilities Index (ECI)Indexes1621620.0

Kimi K3.1 ProvsMuse Wondernormalised score marginsData Nov 15, 2026Data: Vals AI, Arena, Surge AI and 33 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…