Atlas

Compare Models

DeepSeek V4.1 Flash vs Gemini 4 Argon

Gemini 4 Argon leads by 6.5 capability points (≈95% interval +4.3 to +8.7): a clear lead over DeepSeek V4.1 Flash · P(higher) ≈ 100%

Wins 170 of 219 shared benchmarks (34 lost, 15 tied)

Capability vs. relative cost
DeepSeek V4.1 FlashGemini 4 ArgonClosed 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 FlashvsGemini 4 Argoncapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins34 – 170 · 219 shared, 15 tied

Comparing matching setups only (37 benchmarks excluded) · suites counted once
← DeepSeek V4.1 FlashGemini 4 Argon →

DeepSeek V4.1 FlashvsGemini 4 Argonwins by categoryData Nov 15, 2026Data: Vals AI, Artificial Analysis, Arena and 41 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Gemini 4 Argon minus DeepSeek V4.1 Flash.

ARC-AGI-3 (Semi-Private)General QA3.397.5+94.2
ChartographyMultimodal11.671.6+60.0
IOI (Vals v2)Code40.3100.0+59.7
RiemannBenchMath24.076.0+52.0
FrontierMath Tier 4 (v2) - PrivateMath26.878.0+51.2
Concept Synth INDUCTION Challenge64 Round 1General QA17.265.6+48.4
ArXivMath 09/2026 (Native Harnesses)Math40.687.7+47.2
Concept Synth INDUCTION Challenge100 Round 1General QA11.057.0+46.0
ProofBench v1.1Math54.099.0+45.0
Terminal-Bench Science 0.1 (Vals)Science0.044.3+44.3
AA-PersonalityBench Organization indexChat & Writing10056.2−44.2
RareDiseaseBench V1 (35 evaluations)Science0.043.8+43.8
SRE Bench — Fully SolvedCode0.844.3+43.5
Harvey's Legal Agent Benchmark - Capital MarketsLegal0.040.0+40.0
Harvey's Legal Agent Benchmark - Emerging Companies Venture CapitalLegal0.040.0+40.0
Harvey's Legal Agent Benchmark - Funds Asset ManagementLegal0.040.0+40.0
Terminal-Bench-Science 0.1Science9.048.6+39.5
ArXivMath 08/2026 (Native Harnesses)Math43.982.5+38.6
AA-PersonalityBench Transparency indexChat & Writing60.021.9−38.5
AA-PersonalityBench Warmth indexChat & Writing41.079.1+38.5
Furniture Assembly BenchmarkMultimodal34.272.5+38.3
Terminal-Bench 4.0Code19.757.6+37.9
BioSecBench-Refusal V1 Clean RerunsSafety11.348.7+37.4
APEX-Agents 1.1Professional Work50.587.4+36.9
CloudOpsBench v1.0Agents21.857.5+35.8
BuseyBench SVGMultimodal4.17.6+34.7
VIALSMultimodal12.647.2+34.6
ALE-BenchCode18292805+34.6
ArXivLean 09/2026Math5.739.6+34.0
SimpleQA Verified (Epoch anti-abstention)General QA42.976.0+33.1
AA-PersonalityBench Conviction indexChat & Writing96.263.8−32.7
LMArena Agent Arena Chat Net Improvement (Steering Burden)Agents1.614.8+32.1
AA-PersonalityBench Slop indexChat & Writing37.267.6+30.8
Wiz Cyber Model Arena v3.9Cybersecurity47.477.2+29.8
SAFE-Teen 106 worst branch: Health Advice (no prompt)Safety41.170.8−29.7
Decision-Bench v1.1Professional Work25.354.5+29.3
Terminal-Bench 4.0 (Artificial Analysis)Code26.855.1+28.3
SAFE-Teen 106: Health Advice (no prompt)Safety30.257.4−27.3
ExploitGym - release-reported percentage, cohort unspecifiedCybersecurity15.342.0+26.7
Vals Science IndexIndexes29.655.4+25.7
MysteryMechanismScience21.245.5+24.3
TxBench Antibody Discovery V1 (100 evaluations)Science39.763.7+24.0
AA-Omniscience IndexIndexes-5.342.4+23.8
Humanity's Last Exam DiamondGeneral QA19.142.5+23.4
AA-PersonalityBench Expressiveness indexChat & Writing61.984.8+23.1
MineBench — Bradley-Terry RatingGames18222195+22.6
Code MigrationCode45.668.2+22.6
τ³-Banking (Fireworks, banking_knowledge)Agents35.758.2+22.5
SAFE-Teen 106: Urgent Safety (no prompt)Safety11.333.1−21.8
Finance Agent v2 - PrecedentsFinance & Accounting28.249.8+21.6
LLM Chess EloGames10871609+21.4
APEX-Accounting 1.1Finance & Accounting50.171.5+21.4
SAFE-Teen 106 worst branch: Urgent Safety (no prompt)Safety31.452.6−21.2
Mystery Game PuzzlesGames43.064.0+21.0
GameBench 2Games68.489.2+20.7
SEC-Bench Pro - release-reported cohortCybersecurity62.883.4+20.6
Ramp Accounting BenchFinance & Accounting40.861.1+20.3
τ³-Banking (Mercor run, bm25_grep retrieval)Agents37.858.1+20.3
LMArena Agent Arena Work Net Improvement (Steering Burden)Agents2.510.9+20.2
Harvey's Legal Agent Benchmark - Antitrust CompetitionLegal0.020.0+20.0
Harvey's Legal Agent Benchmark - Corporate GovernanceLegal0.020.0+20.0
Harvey's Legal Agent Benchmark - ImmigrationLegal0.020.0+20.0
Harvey's Legal Agent Benchmark - Intellectual PropertyLegal0.020.0+20.0
Harvey's Legal Agent Benchmark - International Trade SanctionsLegal0.020.0+20.0
Harvey's Legal Agent Benchmark - Trusts Estates Private ClientLegal20.040.0+20.0
GBENCH One-shot Coding GScoreGames48.067.9+19.9
Conceptual Reasoning IndexIndexes58.472.4+19.2
VariantBench V1 Clean RerunsScience31.150.3+19.2
TissueMicroscopyBench V1 (49 evaluations)Science2.721.8+19.1
Eval Connections One-Shot with TrapsGames73.092.0+19.0
Humanity's Last ExamGeneral QA36.855.6+18.8
OncologyBench Strict (25 tasks, October 9 rubric)Health & Medicine14.733.3+18.7
FrontierMath Tiers 1-3 (v2) - PrivateMath67.486.0+18.6
ARC-AGI-2General QA72.991.3+18.3
MedXpertQA MM Extended (Mercor held-out)Health & Medicine48.466.7+18.3
Bug Hunt BenchCode20.638.7+18.1
Excel Modeling BenchmarkFinance & Accounting57.275.2+18.0
TxBench Preclinical Pharmacology V1 Clean RerunsScience46.764.7+18.0
MedCodeHealth & Medicine41.258.8+17.6
LMArena Agent Arena Code Net Improvement (Steering Burden)Agents2.710.2+16.9
Terminal-Bench 2.1Code74.557.7−16.9
Code2Video-AA EloCode10531259+16.7
LMCAGeneral QA47.063.7+16.7
TxBench-Antibody-Development V1 (60 evaluations)Science14.431.1+16.7
Tax Agent BenchFinance & Accounting62.579.0+16.6
Finance Agent v2 - Market AnalysisFinance & Accounting63.479.7+16.3
TxBench Oligo Discovery V1 (113 evaluations)Science34.851.0+16.2
SpeciEval (Pooled English Runs)Safety97.481.3−16.1
AA-PersonalityBench Imagination indexChat & Writing31.546.7+15.4
Finance Agent v2 - All PassFinance & Accounting41.155.8+14.7
LMArena Agent Arena Bash RecoveryAgents7.9-21.4−14.6
LLM Creative Story WritingChat & Writing-0.50.8+14.6
SAFE-Teen 106 worst branch: Relationships (no prompt)Safety51.165.7−14.5
WeirdML v3Science5.920.3+14.3
BigFinanceBench private holdout (final-answer accuracy)Finance & Accounting40.654.7+14.1
AA-AnalystAgent (pass^5)Professional Work35.048.8+13.8
ScBench-Long V1 Clean RerunsScience14.828.4+13.6
LMArena Agent Arena Chat Praise vs ComplaintAgents3.130.2+13.6
Legal Research BenchLegal41.354.8+13.5
Legacy-Bench (Fireworks, avg@3)Code57.744.3−13.3
Finance Agent v2 - Disclosure AnalysisFinance & Accounting61.374.5+13.2
Artificial Analysis Intelligence Index v4.3.2Indexes39.552.6+13.1
CWE-bench v1 Programmatic Pass@1Cybersecurity55.068.0+13.0
CritPt (avg@5)Science14.327.1+12.9
DAXBench v2Code75.588.3+12.8
BioSecBench-Structure V1 (58 evaluations)Science27.614.9−12.6
Westworld Finance Diligence BenchFinance & Accounting48.360.7+12.5
BioSecBench-RespiratoryVirus-Surveillance V1 (110 evaluations)Science5.217.6+12.4
Finance Agent v2 - Earnings AnalysisFinance & Accounting72.984.8+11.9
Code Migration v1.1Code58.970.8+11.9
HealthBench (length-adjusted, Artificial Analysis)Health & Medicine60.949.1−11.8
OpenWorldEngineer (Mean@8)Code12.424.1+11.7
Finance Agent v2 - General Quantitative AnalysisFinance & Accounting69.881.5+11.7
HealthBench ProfessionalHealth & Medicine52.964.2+11.3
Finance Agent v2 - AdjustmentsFinance & Accounting49.260.5+11.3
DistressBenchSafety79.868.6−11.1
Creative Writing v3 EloChat & Writing15401758+11.0
Chess Puzzles (task version 2.0)Games48.059.0+11.0
LMArena Agent Arena Work Praise vs ComplaintAgents3.725.1+10.7
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & Medicine23.913.9−10.0
SciCode 1.0.1 (Artificial Analysis)Code51.961.8+10.0
Bedside BenchHealth & Medicine82.992.7+9.8
SWE-bench Multimodal (Vals v1, 480 Tasks)Code46.556.0+9.6
Finance Agent v2 - General Qualitative AnalysisFinance & Accounting67.477.0+9.6
ACCoRDGeneral QA68.177.3+9.2
LMArena Agent Arena Code Praise vs ComplaintAgents3.021.1+9.0
BullshitBench v2 LeaderboardSafety57.066.0+9.0
Vals Finance IndexIndexes57.166.0+8.9
AutomationBench-AA 1.0.6Agents68.977.5+8.6
GDP.pdfProfessional Work19.828.4+8.6
BioMysteryBench (Vals Terminus 2)Science67.876.3+8.5
BioSecBench-Surveillance V1 Clean RerunsScience51.559.8+8.3
LMArena Text Arena EloChat & Writing14751522+8.3
Epoch Capabilities Index (ECI)Indexes156167+8.1
LMArena Code Arena WebDevCode16211682+8.0
DTBenchGeneral QA89.997.6+7.7
Artificial Analysis Progress Index v1Indexes320342+7.6
Creative Writing v3 Rubric ScoreChat & Writing14.115.6+7.5
LMArena Agent Arena Chat Confirmed SuccessAgents4.919.7+7.4
LMArena Agent Arena Chat Steering BurdenAgents-0.514.3+7.4
SAFE-Teen 106: Health Advice (under-18 prompt)Safety23.630.8−7.2
Excel Modeling Benchmark v1.1Finance & Accounting67.874.8+7.1
Longform Creative WritingChat & Writing74.080.9+6.9
DualEntry Accounting AI BenchmarkFinance & Accounting84.291.1+6.9
Finance Agent v2 - Financial ModelingFinance & Accounting23.129.8+6.8
LMArena Agent Arena Code Steering BurdenAgents2.315.7+6.7
LMArena Agent Arena Work Steering BurdenAgents1.514.9+6.7
SAFE-Teen 106 worst branch: Health Advice (under-18 prompt)Safety37.043.6−6.6
Harvey HLAB — Criteria Pass RateLegal87.594.0+6.5
LiveBench 2.0Agents78.084.1+6.1
SAFE-Teen 106 worst branch: Relationships (under-18 prompt)Safety19.814.0+5.8
SpatialBench V1 Clean RerunsScience63.269.0+5.8
SAGEMath47.953.6+5.8
Public Benefits Bench v1.1Professional Work64.369.8+5.5
BrokenArXiv 09/2026 (Native Harnesses)Math54.749.7−5.0
LegalBenchLegal83.388.3+5.0
Senior SWE-BenchCode22.727.7+5.0
Finance Agent v2 - ComparablesFinance & Accounting46.151.0+4.9
Vibe Code Bench 1–100Code16.421.2+4.8
Arena Alignment IndexSafety74.879.4+4.7
ScBench V1 Clean RerunsScience54.759.1+4.5
Access to Justice BenchmarkLegal72.968.5−4.4
MetagenomicsBench V1Science38.042.3+4.3
SAFE-Teen 106: Urgent Safety (under-18 prompt)Safety7.912.1−4.2
CyberBench v1.1 OverallCybersecurity73.777.9+4.2
SAGE v1.1Math85.489.5+4.1
ARC-AGI-1 Semi-PrivateGeneral QA94.598.5+4.0
AA-LCR Extended (Mercor held-out)Search70.874.7+3.9
AA-PersonalityBench Informativeness indexChat & Writing79.182.9+3.8
AA-Briefcase v1.1 EloAgents14151485+3.8
DeepSWE v1.1Code74.277.9+3.7
BioSecBench-Function V1 Clean RerunsScience40.844.4+3.6
ComplexConstraintsChat & Writing49.753.3+3.6
LMArena Agent Arena Code Bash ReliabilityAgents-0.2-7.2−3.5
EpiBench V1 Clean RerunsScience25.528.9+3.5
SAFE-Teen 106 worst branch: Urgent Safety (under-18 prompt)Safety27.230.2−3.0
LMArena Agent Arena Code Confirmed SuccessAgents6.612.3+2.9
LMArena Agent Arena Work Bash ReliabilityAgents0.1-5.7−2.9
NeuroscienceBench V1 (91 evaluations)Science22.019.1−2.9
LMArena Agent Arena Work Confirmed SuccessAgents5.711.4+2.9
FrontierCode 1.1 Extended - Pass RateCode62.064.8+2.8
FrontierCode 1.1 Extended - ScoreCode56.659.4+2.8
APEX-SWECode51.949.2−2.7
LMArena Agent Arena Chat Bash ReliabilityAgents-1.04.2+2.6
Immigration BenchLegal65.668.1+2.4
SpatialBench-Long V1 Clean RerunsScience26.228.6+2.4
MacroscopeBenchCode70.572.8+2.3
Finance Agent v2.1Finance & Accounting65.467.7+2.2
Harvey LAB-AA v1.1 (hallucination-gated)Legal1.73.9+2.2
ProgramBenchCode0.52.5+2.0
MedScribeHealth & Medicine85.587.4+1.9
HealthBenchHealth & Medicine61.960.0−1.8
OTIS Mock AIME 2024-2025Math98.3100.0+1.7
Design Arena Overall Code CategoriesCode13161326+1.7
GDPval-AA v2.1Professional Work16001634+1.5
SAFE-Teen 106: Relationships (no prompt)Safety27.128.5−1.4
SAFE-Teen 106: Relationships (under-18 prompt)Safety4.23.0+1.2
STEMBench120 v1.0.0 (closed-book)Science2.52.4−1.2
GDPval (Mercor 213-task rubric score)Professional Work86.785.9−0.8
BrokenArXiv 08/2026 (Native Harnesses)Math54.854.2−0.6
Financial Coaching BenchFinance & Accounting72.873.1+0.3
LMArena Agent Arena Code Tool HallucinationAgents0.50.3−0.1
LMArena Agent Arena Work Tool HallucinationAgents0.50.3−0.1
LMArena Agent Arena Chat Tool HallucinationAgents0.30.2−0.1
Harvey's Legal Agent Benchmark - Arbitration International Dispute ResolutionLegal0.00.00.0
Harvey's Legal Agent Benchmark - Banking FinanceLegal0.00.00.0
Harvey's Legal Agent Benchmark - Bankruptcy RestructuringLegal0.00.00.0
Harvey's Legal Agent Benchmark - Corporate M&ALegal20.020.00.0
Harvey's Legal Agent Benchmark - Data Privacy CybersecurityLegal20.020.00.0
Harvey's Legal Agent Benchmark - Employment LaborLegal0.00.00.0
Harvey's Legal Agent Benchmark - Energy Natural ResourcesLegal20.020.00.0
Harvey's Legal Agent Benchmark - Environmental ESGLegal0.00.00.0
Harvey's Legal Agent Benchmark - Healthcare Life SciencesLegal20.020.00.0
Harvey's Legal Agent Benchmark - InsuranceLegal0.00.00.0
Harvey's Legal Agent Benchmark - Litigation Dispute ResolutionLegal0.00.00.0
Harvey's Legal Agent Benchmark - Real EstateLegal0.00.00.0
Harvey's Legal Agent Benchmark - Structured Finance SecuritizationLegal0.00.00.0
Harvey's Legal Agent Benchmark - TaxLegal0.00.00.0
Harvey's Legal Agent Benchmark - White Collar Defense InvestigationsLegal0.00.00.0

DeepSeek V4.1 FlashvsGemini 4 Argonnormalised score marginsData Nov 15, 2026Data: Vals AI, Artificial Analysis, Arena and 41 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…