Atlas

Compare Models

GPT-6 Sol vs Muse Wonder

GPT-6 Sol leads by 0.2 capability points (≈95% interval −2.1 to +2.5): statistically tied · P(higher) ≈ 57%

Muse Wonder wins 73 of 125 shared benchmarks (52 lost)

Capability vs. relative cost
GPT-6 SolMuse 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.

GPT-6 SolvsMuse Wondercapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins52 – 73 · 125 shared

Comparing matching setups only (41 benchmarks excluded) · suites counted once
← GPT-6 SolMuse Wonder →

GPT-6 SolvsMuse Wonderwins by categoryData Nov 15, 2026Data: Vals AI, Artificial Analysis, Mercor and 28 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Muse Wonder minus GPT-6 Sol.

AA-PersonalityBench Informativeness indexChat & Writing10.594.3+84.6
AA-PersonalityBench Organization indexChat & Writing12.490.5+78.8
AA-PersonalityBench Imagination indexChat & Writing29.690.5+61.5
LisanBench Path LengthGames822227973+61.4
AA-PersonalityBench Expressiveness indexChat & Writing33.488.6+55.8
LisanBench Difficulty-WeightedGames31188303+55.5
AA-PersonalityBench Slop indexChat & Writing1.054.3+53.8
AA-PersonalityBench Warmth indexChat & Writing54.3100+46.2
VoxelBench Text RatingGames23941870−37.3
DeepResearch Bench II (Mercor 130-task set)Search42.910.8−32.1
Vending-Bench 2Games144289432−30.2
AA-PersonalityBench Conviction indexChat & Writing50.579.1+28.8
EBR-bench v4 (card ban, single agent)Games53.331.4−21.9
Legal Research BenchLegal28.850.5+21.6
Professional Reasoning Benchmark - LegalLegal38.658.4+19.8
FrontierMath Tier 4 (v2) - PrivateMath90.070.7−19.3
Harvey LAB (Mercor run, undisclosed ~213-task cohort)Legal4.123.0+18.9
EnterpriseBench: CoreCraftProfessional Work43.161.5+18.4
Tax Agent BenchFinance & Accounting53.071.2+18.1
DistressBenchSafety68.085.9+17.9
BioSecBench-Refusal V1 Clean RerunsSafety37.955.3+17.4
AA-PersonalityBench Transparency indexChat & Writing39.156.2+17.3
Finance Agent v2 - General Qualitative AnalysisFinance & Accounting53.670.8+17.2
Eval Connections One-Shot with TrapsGames89.073.0−16.0
IOI (Vals v2)Code82.666.9−15.7
Harvey's Legal Agent BenchmarkLegal1.717.1+15.4
ChessBench EloGames13031046−15.3
DAYJOB: HealthcareHealth & Medicine3.618.0+14.4
Public Benefits Bench v1.1Professional Work56.670.2+13.6
Humanity's Sixth SenseMultimodal31.243.7+12.5
Terminal-Bench 4.0 (Artificial Analysis)Code26.338.4+12.1
GBENCH One-shot Coding GScoreGames70.158.1−12.0
DAYJOB: FinanceFinance & Accounting9.321.3+12.0
ChessBench CoherenceGames88.2100.0+11.8
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & Medicine33.922.2−11.7
Finance Agent v2 - Disclosure AnalysisFinance & Accounting55.667.3+11.6
APEX-1 (Mercor, mean score)Professional Work63.675.1+11.5
GDPval Extended (Mercor held-out rubric score)Professional Work58.770.0+11.3
GDPval-AA v2.1Professional Work15141770+11.3
Financial Coaching BenchFinance & Accounting62.673.8+11.2
AA-Briefcase v1.1 EloAgents14761659+10.0
Finance Agent v2 - All PassFinance & Accounting36.045.9+10.0
LMArena Agent Arena Net Improvement (Steering Burden)Agents6.91.8−9.7
CritPt v2 (agentic, avg@5)Science80.971.1−9.7
APEX-SWECode45.054.6+9.6
LMArena Text Arena EloChat & Writing14571511+9.5
Finance Agent v2 - AdjustmentsFinance & Accounting43.252.7+9.4
BuseyBench SVGMultimodal7.56.6−9.3
NoCoT-Bench NCRI 15.2General QA11597.7−9.3
τ³-Banking (Mercor run, bm25_grep retrieval)Agents43.051.9+8.9
MMMU-Pro Extended (Mercor held-out)Multimodal51.943.2−8.7
APEX-Agents 1.1Professional Work68.777.0+8.3
BioMysteryBench (Vals Terminus 2)Science74.867.0−7.8
FrontierMath Tiers 1-3 (v2) - PrivateMath89.882.1−7.7
ChessBench AccuracyGames77.470.2−7.7
FrontierCode 2.0 (Artificial Analysis)Code35.627.9−7.7
Toloka Arena (v3)Agents58.065.7+7.7
Finance Agent v2 - General Quantitative AnalysisFinance & Accounting70.077.6+7.6
APEX-Accounting 1.1Finance & Accounting59.767.1+7.4
Terminal-Bench Science 0.1 (Vals)Science30.022.9−7.1
Finance Agent v2 - ComparablesFinance & Accounting40.847.9+7.1
DualEntry Accounting AI BenchmarkFinance & Accounting76.283.2+7.0
NoCoT-Bench NCKI (kspine v3)General QA119109−6.4
AutomationBench-AA 1.0.6Agents61.768.1+6.4
Harvey LAB-AA v1.1 (hallucination-gated)Legal3.610.0+6.4
HardGeoBench Country AccuracyMultimodal46.352.5+6.3
ACCoRDGeneral QA82.476.2−6.2
Humanity's Last Exam DiamondGeneral QA45.239.0−6.2
Creative Writing v3 EloChat & Writing21262007−6.0
TaxCalcBench TY25Finance & Accounting28.022.0−6.0
LMArena Agent Arena Praise vs ComplaintAgents10.0-1.8−5.9
Vals Science IndexIndexes45.039.1−5.9
Code MigrationCode57.262.6+5.4
Terminal-Bench 4.0Code44.449.5+5.1
Mystery Game PuzzlesGames56.051.0−5.0
Artificial Analysis Intelligence Index v4.3.2Indexes47.652.4+4.8
Vibe Code Bench v1.1Code87.883.1−4.7
WeirdML v3Science19.715.2−4.6
Conceptual Reasoning IndexIndexes71.067.7−4.5
Excel Modeling BenchmarkFinance & Accounting71.567.1−4.4
Finance Agent v2 - Earnings AnalysisFinance & Accounting72.476.8+4.4
SpatialBench V1 Clean RerunsScience67.571.9+4.3
Terminal-Bench-Science 0.1Science30.025.7−4.3
AA-LCR Extended (Mercor held-out)Search70.274.4+4.2
CharXiv Extended (Mercor held-out)Multimodal83.879.6−4.2
Furniture Assembly BenchmarkMultimodal58.354.2−4.2
Finance Agent v2 - Market AnalysisFinance & Accounting58.062.1+4.2
OncologyBench Strict (25 tasks, October 9 rubric)Health & Medicine20.024.0+4.0
HardGeoBench Avg DistanceMultimodal14971369+4.0
LiveBench 2.0Agents81.978.1−3.8
Humanity's Last Exam (Text-Only)General QA46.350.1+3.8
LMArena Agent Arena Steering BurdenAgents10.12.6−3.8
Finance Agent v2 - PrecedentsFinance & Accounting28.532.1+3.6
Vals Index v2.1Indexes57.561.1+3.5
GDPval (Mercor 213-task rubric score)Professional Work83.587.0+3.5
Finance Agent v2 - Financial ModelingFinance & Accounting19.322.5+3.2
SciCode (main problems)Science12.715.8+3.1
LMArena Agent Arena Bash ReliabilityAgents10.94.8−3.1
BullshitBench v2 LeaderboardSafety54.057.0+3.0
GDP.pdfProfessional Work28.030.8+2.8
MysteryMechanismScience30.227.5−2.7
SciCode 1.0.1 (Artificial Analysis)Code57.660.3+2.7
CritPt (avg@5)Science30.928.3−2.6
MacroscopeBenchCode61.864.2+2.4
CyberBench v1.1 OverallCybersecurity78.080.4+2.4
MMMU ProMultimodal82.584.9+2.4
LMCAGeneral QA59.156.8−2.2
DTBenchGeneral QA97.395.2−2.1
Decision-Bench v1.1Professional Work36.534.6−1.9
GameBench 2Games79.477.8−1.6
Epoch Capabilities Index (ECI)Indexes164162−1.6
HardGeoBench <=10miMultimodal26.327.5+1.3
HardGeoBench <=100miMultimodal37.538.8+1.3
ProofBench v1.1Math83.084.0+1.0
WellnessBenchSafety56.055.0−1.0
SpeciEval (Pooled English Runs)Safety98.597.5−1.0
Creative Writing v3 Rubric ScoreChat & Writing16.116.2+0.8
AA-Omniscience IndexIndexes27.125.5−0.8
LMArena Agent Arena Tool HallucinationAgents1.0-0.5−0.8
MMLU ProGeneral QA88.989.6+0.7
Artificial Analysis Progress Index v1Indexes334336+0.6
Artificial Analysis Intelligence Index v5.0.1Indexes44.243.6−0.6
OTIS Mock AIME 2024-2025Math100.099.4−0.6
CharXiv (descriptive and reasoning combined)Multimodal95.094.5−0.5
LMArena Agent Arena Confirmed SuccessAgents4.55.4+0.4

GPT-6 SolvsMuse Wondernormalised score marginsData Nov 15, 2026Data: Vals AI, Artificial Analysis, Mercor and 28 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…