Atlas

Compare Models

Claude Sonnet 5.5 vs Muse Spark 1.3

Claude Sonnet 5.5 leads by 5.4 capability points (≈95% interval +3.1 to +7.8): a clear lead over Muse Spark 1.3 · P(higher) ≈ 100%

Wins 116 of 164 shared benchmarks (45 lost, 3 tied)

Capability vs. relative cost
Claude Sonnet 5.5Muse Spark 1.3Closed 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 Sonnet 5.5vsMuse Spark 1.3capability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins116 – 45 · 164 shared, 3 tied

Comparing matching setups only (7 benchmarks excluded) · suites counted once
← Claude Sonnet 5.5Muse Spark 1.3 →

Claude Sonnet 5.5vsMuse Spark 1.3wins by categoryData Nov 15, 2026Data: Vals AI, Artificial Analysis, Mercor and 46 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Muse Spark 1.3 minus Claude Sonnet 5.5.

KernelBench Mega Normalized (RTX PRO 6000)Code90.66.2−84.5
AA-PersonalityBench Transparency indexChat & Writing10041.0−59.6
KernelBench CUDA Normalized (RTX PRO 6000)Code87.030.3−56.7
InferenceBenchCode9.14.5−54.1
AA-PersonalityBench Slop indexChat & Writing27.775.3+48.1
Terminal-Bench-Science 0.1Science53.311.0−42.4
ProofBench v1.1Math100.058.0−42.0
HoneyBench v0.2 (hack rate)Safety41.683.6−42.0
Mystery Game PuzzlesGames65.025.0−40.0
Terminal-Bench Science 0.1 (Vals)Science47.110.0−37.1
FrontierSWE V2Code61.925.8−36.1
Olam Social Poker EloGames15261494−34.8
FrontierML ($200 wallet)Agents50.416.0−34.5
Code2Video-AA EloCode14431019−34.5
FrontierMath Tier 4 (v2) - PrivateMath80.546.3−34.1
Terminal-Bench 4.0Code61.827.8−34.0
ChartographyMultimodal61.627.6−34.0
MatFlowBench Pass@1Science50.016.7−33.3
Medical Long Context Reasoning (MLCR-AA)Health & Medicine75.043.3−31.7
Decision-Bench v1.1Professional Work51.219.8−31.4
Terminal-Bench 4.0 (Artificial Analysis)Code63.633.3−30.3
DAYJOB: Data ScienceProfessional Work35.36.0−29.3
SAFE-Teen 106 worst branch: Health Advice (no prompt)Safety42.913.9+29.0
sudo L7Code41.713.9−27.8
APEX-SWECode66.439.1−27.3
ChessBench CoherenceGames64.691.2+26.6
IOI (Vals v2)Code83.156.6−26.5
SAFE-Teen 106: Health Advice (no prompt)Safety33.26.7+26.5
SAFE-Teen 106 worst branch: Health Advice (under-18 prompt)Safety30.84.5+26.3
LMArena Agent Arena Net Improvement (Steering Burden)Agents16.02.2−26.0
SAFE-Teen Health Advice (no prompt)Safety30.24.6+25.6
AA-3D v1.0 EloCode12041032−25.0
Arena Image-to-WebDevCode17441637−23.3
RiemannBench (November 2026 rescoring)Math76.853.6−23.2
AA-PersonalityBench Warmth indexChat & Writing73.396.2+23.1
GameBench 2Games91.468.5−22.9
Code MigrationCode69.847.4−22.4
APEX-Agents 1.1Professional Work83.761.3−22.4
Chess Puzzles (task version 2.0)Games59.037.0−22.0
AA-PersonalityBench Organization indexChat & Writing84.863.8−21.2
Harvey's Legal Agent BenchmarkLegal2.923.8+20.8
ChessBench EloGames8571201+20.5
AA-AnalystAgent (pass^5)Professional Work57.537.5−20.0
SAFE-Teen Relationships (no prompt)Safety16.035.7−19.7
Humanity's Last Exam DiamondGeneral QA53.333.7−19.6
SAFE-Teen 106 worst branch: Relationships (no prompt)Safety37.056.2−19.2
AA-PersonalityBench Conviction indexChat & Writing58.177.2+19.2
APEX-1 (Mercor, mean score)Professional Work53.973.0+19.1
FrontierCode 2.0 (Artificial Analysis)Code41.723.1−18.6
Bug Hunt BenchCode48.930.7−18.2
Artificial Analysis Intelligence Index v5.0.1Indexes54.536.6−17.9
Artificial Analysis Intelligence Index v5.0Indexes54.236.4−17.8
DeepFinanceFinance & Accounting54.637.0−17.6
SAFE-Teen 106: Relationships (no prompt)Safety16.934.4−17.5
DAYJOB: AccountingFinance & Accounting35.218.0−17.2
ALE-BenchCode24481981−16.6
AA-PersonalityBench Informativeness indexChat & Writing84.8100+15.4
LMArena Code Arena WebDevCode17731655−15.4
SWE-bench Multimodal (Vals v1, 480 Tasks)Code58.343.1−15.2
MineBench — Bradley-Terry RatingGames19981747−15.2
GDPval Extended (Mercor held-out rubric score)Professional Work70.955.9−15.0
Conceptual Reasoning IndexIndexes77.166.2−15.0
SAFE-Teen 106: Health Advice (under-18 prompt)Safety16.92.1+14.8
GBENCH One-shot Coding GScoreGames67.452.7−14.7
LMArena Agent Arena Praise vs ComplaintAgents30.51.7−14.4
LMCAGeneral QA68.353.9−14.4
FrontierMath Tiers 1-3 (v2) - PrivateMath88.874.4−14.4
Artificial Analysis Coding Agent Index v1.5Indexes68.454.3−14.1
AutomationBench-AA 1.0.6Agents71.857.9−13.9
CursorBench 4.0Code55.541.6−13.9
SAFE-Teen Health Advice (under-18 prompt)Safety14.70.8+13.9
HANDBOOK.md Agents v1.1Professional Work47.733.9−13.8
AA-PersonalityBench Imagination indexChat & Writing84.898.1+13.5
MysteryMechanismScience49.136.0−13.1
GeneBench-Pro Public Release SubsetScience39.026.0−13.0
AA-Briefcase v1.1 EloAgents18161581−12.9
TaxCalcBench TY25Finance & Accounting18.06.0−12.0
ACCoRDGeneral QA85.775.1−10.6
DistressBenchSafety74.684.9+10.3
HardGeoBench Avg DistanceMultimodal6891022−10.3
DAYJOB: FinanceFinance & Accounting24.014.8−9.2
Artificial Analysis Progress Index v1Indexes351324−9.2
EnterpriseBench: CoreCraftProfessional Work70.861.9−8.9
Excel Modeling BenchmarkFinance & Accounting75.767.4−8.3
BuseyBench SVGMultimodal6.55.7−8.2
Eval Connections One-Shot with TrapsGames76.084.0+8.0
Artificial Analysis Intelligence Index v4.3.2Indexes56.048.1−7.9
LMArena Agent Arena Confirmed SuccessAgents19.53.7−7.9
SpeciEval (Refusal-Aware Retries, Historical)Safety91.098.8+7.8
SpeciEval (Pooled English Runs)Safety91.098.7+7.8
CritPt v2 (agentic, avg@5)Science75.768.0−7.7
LiveBench 2.0Agents82.574.8−7.7
DAYJOB: HealthcareHealth & Medicine15.27.8−7.4
Vision Arena OverallMultimodal12631291+7.4
DAYJOB: TaxFinance & Accounting10.83.6−7.2
GDPval-AA v2.1Professional Work18381681−6.9
AA-LCR Extended (Mercor held-out)Search73.866.9−6.9
Epoch Capabilities Index (ECI)Indexes166158−6.8
Vibe Code Bench v1.1Code92.485.9−6.5
GDP.xlsxProfessional Work29.122.6−6.5
CVDP (517-problem set)Code60.654.3−6.3
Harvey LAB-AA v1.1 (hallucination-gated)Legal2.88.9+6.1
LMArena Agent Arena Steering BurdenAgents14.52.6−5.9
SimpleBenchGeneral QA75.981.8+5.9
SAFE-Teen 106: Urgent Safety (no prompt)Safety14.79.0+5.7
CritPt (avg@5)Science31.426.0−5.4
LMArena Agent Arena Bash ReliabilityAgents12.72.0−5.4
FrontierCode 1.1 Main (Artificial Analysis)Code45.840.6−5.2
SAFE-Teen Urgent Safety (no prompt)Safety14.49.3+5.1
SWE-Marathon v1.1Code51.346.3−5.0
Extended NYT Connections 940 (Quadratic v1)Games80.585.1+4.6
MedXpertQA MM Extended (Mercor held-out)Health & Medicine55.660.1+4.5
LMArena Agent Arena Bash RecoveryAgents15.06.1−4.5
MedXpertQA MM (200-question set)Health & Medicine80.776.3−4.4
DAYJOB: LawLegal27.022.8−4.2
LMArena Text Arena EloChat & Writing14721496+4.2
ATLAS (Exa, 200-task subset)Search46.842.7−4.1
SAFE-Teen 106 worst branch: Relationships (under-18 prompt)Safety7.33.2+4.1
WellnessBenchSafety57.061.0+4.0
ProgramBenchCode6.52.5−4.0
Roboflow Vision Evals (September 2026 protocol)Multimodal83.879.8−4.0
Immigration BenchLegal69.065.0−4.0
Financial Coaching BenchFinance & Accounting74.971.0−4.0
NoCoT-Bench NCKI (kspine v3)General QA107113+3.9
GDP.pdfProfessional Work31.427.6−3.8
AA-Omniscience IndexIndexes32.325.0−3.7
RuneBenchGames4.84.6−3.6
AA-LCR v1.1Search79.783.0+3.3
SWE-bench Verified Extended (Mercor held-out)Code57.354.0−3.3
Humanity's Last ExamGeneral QA46.143.2−2.9
Access to Justice BenchmarkLegal74.972.2−2.8
DeepSWE v1.1Code70.567.8−2.7
micro1 PharmaBenchHealth & Medicine11.413.9+2.5
HardGeoBench Country AccuracyMultimodal62.560.0−2.5
CharXiv Extended (Mercor held-out)Multimodal84.281.8−2.4
ComplexConstraints (November 2026 rescoring)Chat & Writing67.264.8−2.4
SAFE-Teen 106 worst branch: Urgent Safety (no prompt)Safety23.225.6−2.4
CharXiv (descriptive and reasoning combined)Multimodal96.193.9−2.2
AA-PersonalityBench Expressiveness indexChat & Writing73.371.4−1.9
τ³-Banking (Mercor run, bm25_grep retrieval)Agents54.052.2−1.8
NoCoT-Bench NCRI 15.2General QA95.198.3+1.8
Terminal-Bench 2.1Code87.388.8+1.5
SciCode 1.0.1 (Artificial Analysis)Code61.059.7−1.3
HardGeoBench <=100miMultimodal53.855.0+1.3
HardGeoBench <=10miMultimodal40.038.8−1.2
SAFE-Teen Relationships (under-18 prompt)Safety1.80.6+1.1
DAXBench v2Code86.285.1−1.1
SAFE-Teen 106: Relationships (under-18 prompt)Safety1.70.6+1.1
MazeBench ASCIIGames1.00.0−1.0
OTIS Mock AIME 2024-2025Math100.099.2−0.8
DTBenchGeneral QA97.396.5−0.8
ForecastBench Preliminary Dataset Brier IndexForecasting61.161.9+0.8
ObviousBenchGeneral QA100.099.3−0.7
SAFE-Teen 106 worst branch: Urgent Safety (under-18 prompt)Safety11.811.3+0.5
APEX-Accounting 1.1Finance & Accounting64.564.9+0.4
Ruby LLM Benchmarks — Program FixerCode73.874.1+0.3
SAFE-Teen 106: Urgent Safety (under-18 prompt)Safety2.92.7+0.2
ChessBench AccuracyGames75.876.0+0.2
SAFE-Teen Urgent Safety (under-18 prompt)Safety2.92.7+0.2
LMArena Agent Arena Tool HallucinationAgents0.90.8−0.1
Ramp Accounting BenchFinance & Accounting42.542.50.0
AIME 2025Math100.0100.00.0
SWE-TogetherCode62.462.40.0
Visa Agentic Commerce BenchmarkAgents64.464.40.0

Claude Sonnet 5.5vsMuse Spark 1.3normalised score marginsData Nov 15, 2026Data: Vals AI, Artificial Analysis, Mercor and 46 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…