Atlas

Compare Models

DeepSeek V4.1 Flash vs Muse Wonder Full

Muse Wonder Full leads by 5.3 capability points (≈95% interval +3.1 to +7.5): a clear lead over DeepSeek V4.1 Flash · P(higher) ≈ 100%

Wins 68 of 74 shared benchmarks (5 lost, 1 tied)

Capability vs. relative cost
DeepSeek V4.1 FlashMuse 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.

DeepSeek V4.1 FlashvsMuse Wonder Fullcapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins5 – 68 · 74 shared, 1 tied

Comparing matching setups only (6 benchmarks excluded) · suites counted once
← DeepSeek V4.1 FlashMuse Wonder Full →

DeepSeek V4.1 FlashvsMuse 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 DeepSeek V4.1 Flash.

ProofBench v1.1Math54.094.0+40.0
IOI (Vals v2)Code40.379.8+39.6
Terminal-Bench 4.0Code19.758.1+38.4
Terminal-Bench Science 0.1 (Vals)Science0.037.1+37.1
Humanity's Last Exam DiamondGeneral QA19.148.5+29.4
Terminal-Bench-Science 0.1Science9.037.1+28.1
LMArena Agent Arena Chat Net Improvement (Steering Burden)Agents1.610.8+22.5
GDP.pdf (Artificial Analysis, All-pass)Professional Work12.832.6+19.8
LMArena Agent Arena Code Net Improvement (Steering Burden)Agents2.711.4+19.5
Bug Hunt BenchCode20.640.0+19.4
Vals Science IndexIndexes29.648.5+18.9
CritPt (avg@5)Science14.332.9+18.6
AA-Omniscience IndexIndexes-5.330.1+17.7
LMArena Agent Arena Work Net Improvement (Steering Burden)Agents2.59.8+17.6
AA-Briefcase v1.1 EloAgents14151729+17.2
Artificial Analysis Intelligence Index v4.3.2Indexes39.555.8+16.3
Code MigrationCode45.661.0+15.4
Harvey's Legal Agent BenchmarkLegal6.721.7+15.0
Excel Modeling BenchmarkFinance & Accounting57.271.9+14.7
MysteryMechanismScience21.235.1+14.0
LMArena Code Arena WebDevCode16211723+13.5
Tax Agent BenchFinance & Accounting62.575.0+12.5
Legal Research BenchLegal41.353.4+12.0
Harvey LAB-AA v1.1 (hallucination-gated)Legal1.713.1+11.4
GDPval-AA v2.1Professional Work16001857+11.3
SRE Bench — Fully SolvedCode0.811.8+11.1
SciCode 1.0.1 (Artificial Analysis)Code51.962.6+10.8
Finance Agent v2 - General Quantitative AnalysisFinance & Accounting69.880.1+10.2
Finance Agent v2 - Disclosure AnalysisFinance & Accounting61.371.6+10.2
Public Benefits Bench v1.1Professional Work64.373.7+9.5
LMArena Agent Arena Chat Praise vs ComplaintAgents3.121.3+9.1
MedCodeHealth & Medicine41.249.9+8.8
Artificial Analysis Progress Index v1Indexes320343+8.1
Finance Agent v2 - All PassFinance & Accounting41.149.0+8.0
Vals Finance IndexIndexes57.164.6+7.5
Finance Agent v2 - AdjustmentsFinance & Accounting49.256.7+7.4
LMArena Agent Arena Code Praise vs ComplaintAgents3.017.6+7.3
Excel Modeling Benchmark v1.1Finance & Accounting67.875.0+7.2
Finance Agent v2 - PrecedentsFinance & Accounting28.235.0+6.8
Finance Agent v2 - General Qualitative AnalysisFinance & Accounting67.474.2+6.8
LMArena Agent Arena Work Praise vs ComplaintAgents3.716.5+6.4
LMArena Agent Arena Code Steering BurdenAgents2.314.9+6.3
Finance Agent v2 - Earnings AnalysisFinance & Accounting72.978.6+5.7
Immigration BenchLegal65.671.2+5.6
BioMysteryBench (Vals Terminus 2)Science67.873.3+5.6
LMArena Agent Arena Chat Bash ReliabilityAgents-1.09.8+5.4
LMArena Agent Arena Work Steering BurdenAgents1.512.2+5.3
DeepSWE v1.1Code74.279.0+4.8
LMArena Agent Arena Chat Confirmed SuccessAgents4.914.1+4.6
LMArena Agent Arena Chat Steering BurdenAgents-0.58.2+4.4
LegalBenchLegal83.387.4+4.1
LMArena Agent Arena Work Bash ReliabilityAgents0.18.2+4.0
LMArena Agent Arena Code Bash ReliabilityAgents-0.27.9+4.0
LMArena Agent Arena Code Confirmed SuccessAgents6.614.3+3.9
Code Migration v1.1Code58.962.8+3.9
MedScribeHealth & Medicine85.589.3+3.8
LMArena Agent Arena Work Confirmed SuccessAgents5.710.9+2.6
AutomationBench-AA 1.0.6Agents68.971.3+2.4
Finance Agent v2.1Finance & Accounting65.467.6+2.1
Financial Coaching BenchFinance & Accounting72.875.0+2.1
Access to Justice BenchmarkLegal72.974.7+1.8
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & Medicine23.922.2−1.7
Finance Agent v2 - Market AnalysisFinance & Accounting63.464.9+1.6
Vibe Code Bench v1.1Code84.786.0+1.3
FrontierCode 1.1 Main (Artificial Analysis)Code44.044.9+0.9
SWE-bench Multimodal (Vals v1, 480 Tasks)Code46.547.3+0.8
Finance Agent v2 - ComparablesFinance & Accounting46.146.8+0.8
Finance Agent v2 - Financial ModelingFinance & Accounting23.123.7+0.7
CyberGymCybersecurity88.188.4+0.3
LMArena Agent Arena Code Tool HallucinationAgents0.50.3−0.1
LMArena Agent Arena Work Tool HallucinationAgents0.50.4−0.1
CyberBench v1.1 OverallCybersecurity73.773.6−0.1
LMArena Agent Arena Chat Tool HallucinationAgents0.30.30.0
ProgramBenchCode0.50.50.0

DeepSeek V4.1 FlashvsMuse 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…