Atlas

Compare Models

Claude Haiku 5.5 vs Gemini 4 Argon

Gemini 4 Argon leads by 4.6 capability points (≈95% interval +2.3 to +6.9): a clear lead over Claude Haiku 5.5 · P(higher) ≈ 100%

Wins 155 of 190 shared benchmarks (23 lost, 12 tied)

Capability vs. relative cost
Claude Haiku 5.5Gemini 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.

Claude Haiku 5.5vsGemini 4 Argoncapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins23 – 155 · 190 shared, 12 tied

Comparing matching setups only (13 benchmarks excluded) · suites counted once
← Claude Haiku 5.5Gemini 4 Argon →

Claude Haiku 5.5vsGemini 4 Argonwins by categoryData Nov 15, 2026Data: Vals AI, Artificial Analysis, Mercor and 45 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 Claude Haiku 5.5.

ARC-AGI-3 (Semi-Private)General QA2.897.5+94.6
AA-PersonalityBench Slop indexChat & Writing10.567.6+57.7
AA-PersonalityBench Transparency indexChat & Writing77.221.9−55.8
IOI (Vals v2)Code47.3100.0+52.7
SimpleQA Verified (Epoch anti-abstention)General QA23.876.0+52.2
TaxCalcBench TY25 (web search)Finance & Accounting22.074.0+52.0
Toloka Arena (v3)Agents21.073.0+52.0
NoCoT-Bench NCRI 15.2General QA68.3151+45.5
BrowseComp Extended (Mercor held-out)Search27.572.0+44.5
AutomationBench-AA 1.0.6Agents36.077.5+41.6
Terminal-Bench-Science 0.1Science7.148.6+41.4
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
Harvey's Legal Agent Benchmark - Trusts Estates Private ClientLegal0.040.0+40.0
RuneBenchGames4.36.9+40.0
Terminal-Bench 4.0 (Artificial Analysis)Code15.255.1+39.9
SRE Bench — Fully SolvedCode5.744.3+38.6
AutomationBench 1.0.6 (Zapier)Agents12.050.1+38.1
MineBench — Bradley-Terry RatingGames15822195+37.1
ChessBench EloGames8751485+36.3
Terminal-Bench Science 0.1 (Vals)Science10.044.3+34.3
Mystery Game PuzzlesGames30.064.0+34.0
FrontierMath Tier 4 (v2) - PrivateMath46.378.0+31.7
BuseyBench SVGMultimodal4.67.6+30.0
ARC-AGI-2General QA61.391.3+30.0
NoCoT-Bench NCKI (kspine v3)General QA85.9128+29.4
Decision-Bench v1.1Professional Work25.354.5+29.3
ChessBenchGames41.370.0+28.7
Terminal-Bench 2.1Code85.057.7−27.3
ALE-BenchCode19962747+26.6
Eval Connections One-Shot with TrapsGames67.092.0+25.0
Furniture Assembly BenchmarkMultimodal47.572.5+25.0
Code MigrationCode44.668.2+23.6
HLAB-Vals-1.1Legal4.227.1+22.9
Finance Agent v2 - PrecedentsFinance & Accounting27.049.8+22.9
Harvey LAB (Mercor run, undisclosed ~213-task cohort)Legal4.927.4+22.5
AA-3D v1.0 EloCode9981147+21.6
Blueprint-Bench 2Multimodal0.30.5+21.4
Terminal-Bench 4.0Code36.457.6+21.2
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & Medicine35.013.9−21.1
EBR-bench v4 (card ban, single agent)Games43.864.8+21.0
Harvey HLAB — Criteria Pass RateLegal73.494.0+20.6
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 - Corporate M&ALegal0.020.0+20.0
Harvey's Legal Agent Benchmark - Data Privacy CybersecurityLegal0.020.0+20.0
Harvey's Legal Agent Benchmark - Energy Natural ResourcesLegal0.020.0+20.0
Harvey's Legal Agent Benchmark - Healthcare Life SciencesLegal0.020.0+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
LMArena Agent Arena Net Improvement (Steering Burden)Agents0.510.7+19.4
LLM Chess EloGames11431609+19.2
APEX-Agents 1.1Professional Work68.387.4+19.1
Ramp Accounting BenchFinance & Accounting42.361.1+18.8
Code Migration v1.1Code52.470.8+18.4
MedXpertQA MM Extended (Mercor held-out)Health & Medicine48.466.7+18.3
Bug Hunt BenchCode20.538.7+18.3
HardGeoBench Avg DistanceMultimodal1319741+17.8
LMArena Text Arena EloChat & Writing14221522+17.5
AA-PersonalityBench Imagination indexChat & Writing63.846.7−17.3
Vals Science IndexIndexes38.555.4+16.9
Tax Agent BenchFinance & Accounting62.279.0+16.8
Finance Agent v2 - Market AnalysisFinance & Accounting63.179.7+16.6
LisanBench Path LengthGames128977560−16.6
Excel Modeling BenchmarkFinance & Accounting59.075.2+16.2
ChessBench AccuracyGames66.981.8+16.0
AA-Omniscience IndexIndexes10.742.4+15.8
AA-PersonalityBench Organization indexChat & Writing41.056.2+15.4
τ³-Banking (Mercor run, bm25_grep retrieval)Agents43.058.1+15.1
Finance Agent v2 - General Qualitative AnalysisFinance & Accounting61.977.0+15.1
HardGeoBench <=10miMultimodal30.045.0+15.0
DualEntry Accounting AI BenchmarkFinance & Accounting76.291.1+14.9
Humanity's Last Exam (Text-Only)General QA42.757.1+14.4
GameBench 2Games75.189.2+14.1
Artificial Analysis Intelligence Index v5.0.1Indexes36.550.4+13.9
LMArena Agent Arena Praise vs ComplaintAgents-3.524.1+13.8
Wiz Cyber Model Arena v3.9Cybersecurity57.070.8+13.8
HardGeoBench <=100miMultimodal45.058.8+13.8
Finance Agent v2 - All PassFinance & Accounting42.955.8+12.9
Humanity's Last Exam DiamondGeneral QA29.642.5+12.9
LMCAGeneral QA51.163.7+12.6
HardGeoBench Country AccuracyMultimodal51.363.8+12.5
GBENCH One-shot Coding GScoreGames55.667.9+12.2
LisanBench Difficulty-WeightedGames27113849+12.2
SWE-bench Multimodal (Vals v1, 480 Tasks)Code44.056.0+12.1
SWE-Bench Pro V2 — FullCode87.299.2+12.0
ChessBench CoherenceGames80.792.6+11.9
FrontierSWE V2 Best@5Code52.864.3+11.6
LMArena Code Arena WebDevCode15941682+11.5
Finance Agent v2 - Earnings AnalysisFinance & Accounting73.584.8+11.3
Legal Research BenchLegal43.854.8+11.1
MultiChallengeChat & Writing67.778.7+11.0
Finance Agent v2 - AdjustmentsFinance & Accounting49.660.5+11.0
AA-LCR Extended (Mercor held-out)Search63.874.7+10.9
FrontierMath Tiers 1-3 (v2) - PrivateMath75.186.0+10.9
SWE Atlas v1.1 - OverallCode59.470.1+10.7
VIALS-AA (pass^3)Multimodal49.159.6+10.6
GDPval Extended (Mercor held-out rubric score)Professional Work66.076.4+10.4
MysteryMechanismScience35.145.5+10.4
MKA (Artificial Analysis)Professional Work41.051.3+10.3
CritPt v2 (agentic, avg@5)Science62.672.9+10.3
Finance Agent v2 - Disclosure AnalysisFinance & Accounting64.374.5+10.2
ProofBench v1.1Math89.099.0+10.0
FrontierSWE V2 Worst@5Code34.544.5+10.0
Conceptual Reasoning IndexIndexes65.372.4+9.8
AA-PersonalityBench Warmth indexChat & Writing69.579.1+9.6
ARC-AGI-1 Semi-PrivateGeneral QA89.098.5+9.5
HealthBench (length-adjusted, Artificial Analysis)Health & Medicine58.549.1−9.4
Excel Modeling Benchmark v1.1Finance & Accounting65.574.8+9.4
APEX-Accounting 1.1Finance & Accounting62.271.5+9.3
GDPval-AA v2.1Professional Work14241634+9.2
MMMU ProMultimodal79.188.3+9.2
Artificial Analysis Intelligence Index v4.3.2Indexes43.452.6+9.2
BullshitBench v2 LeaderboardSafety57.066.0+9.0
WeirdML v3Science11.720.3+8.6
Vals Finance IndexIndexes57.466.0+8.6
LiveBench 2.0Agents75.684.1+8.5
SAGEMath45.853.6+7.9
AA-PersonalityBench Informativeness indexChat & Writing75.382.9+7.7
AA-PersonalityBench Expressiveness indexChat & Writing77.284.8+7.7
Design Arena Overall Code CategoriesCode12801326+7.6
MedCodeHealth & Medicine51.558.8+7.3
Finance Agent v2 - General Quantitative AnalysisFinance & Accounting74.481.5+7.2
SciCode 1.0.1 (Artificial Analysis)Code55.061.8+6.8
Public Benefits Bench v1.1Professional Work63.269.8+6.6
Vals Legal IndexIndexes70.376.8+6.5
LMArena Agent Arena Bash ReliabilityAgents7.4-5.4−6.4
DAXBench v2Code81.988.3+6.4
Immigration BenchLegal61.868.1+6.3
SWE-Marathon v1.1Code50.043.8−6.2
BioMysteryBench (Vals Terminus 2)Science70.476.3+5.9
APEX-SWECode55.149.2−5.9
LMArena Agent Arena Confirmed SuccessAgents1.613.2+5.8
VIALS-AA (average of 3)Multimodal60.566.3+5.8
SpeciEval (Pooled English Runs)Safety86.981.3−5.7
Finance Agent v2 - ComparablesFinance & Accounting45.351.0+5.6
Artificial Analysis Progress Index v1Indexes326342+5.5
Code2Video-AA EloCode11921259+5.4
BioSimBench 1.0Science13.018.4+5.4
LMArena Agent Arena Steering BurdenAgents2.413.2+5.4
Epoch Capabilities Index (ECI)Indexes160167+5.1
AA-AnalystAgent (pass^5)Professional Work53.848.8−5.0
GDP.pdf (Artificial Analysis, All-pass)Professional Work20.825.6+4.8
o11y-bench 1.1 Pass^3Agents85.981.3−4.7
FrontierML ($200 wallet)Agents34.939.5+4.6
CritPt (avg@5)Science22.627.1+4.6
Agent Initiative BenchmarkAgents6.05.6−4.5
CursorBench 4.0Code48.452.9+4.5
ACCoRDGeneral QA81.577.3−4.2
FrontierPhysics v0.1Science4.28.3+4.1
Chess Puzzles (task version 2.0)Games55.059.0+4.0
AA-PersonalityBench Conviction indexChat & Writing67.663.8−3.8
DistressBenchSafety71.768.6−3.0
DTBenchGeneral QA95.097.6+2.6
AA-Briefcase v1.1 EloAgents14411485+2.4
GDPxProfessional Work4.97.3+2.4
CyberBench v1.1 OverallCybersecurity75.577.9+2.4
Finance Agent v2.1Finance & Accounting65.467.7+2.3
FrontierCode 1.1 Extended - ScoreCode58.460.6+2.1
MedScribeHealth & Medicine89.587.4−2.1
Finance Agent v2 - Financial ModelingFinance & Accounting28.329.8+1.5
FrontierCode 1.1 Extended - Pass RateCode64.766.1+1.4
DeepFinanceFinance & Accounting40.041.4+1.4
AA-LCR v1.1Search78.379.7+1.3
FrontierCode 2.0 - ScoreCode28.927.6−1.3
LMArena Agent Arena Tool HallucinationAgents-2.10.5+1.3
FrontierCode 2.0 - Pass RateCode31.930.6−1.2
VIALS-AA (pass@3)Multimodal71.472.7+1.2
HealthBenchHealth & Medicine61.260.0−1.2
OTIS Mock AIME 2024-2025Math98.9100.0+1.1
Vibe Code Bench 1–100Code20.421.2+0.8
HealthBench ProfessionalHealth & Medicine64.864.2−0.5
GDPval (Mercor 213-task rubric score)Professional Work85.585.9+0.4
SAGE v1.1Math89.889.5−0.4
Prophet Arena Arena Score (Agentic, OpenCode)Forecasting85.986.0+0.2
Prophet Arena Arena Score (Fixed Context)Forecasting85.485.3−0.1
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 - Employment LaborLegal0.00.00.0
Harvey's Legal Agent Benchmark - Environmental ESGLegal0.00.00.0
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
ProgramBenchCode2.52.50.0

Claude Haiku 5.5vsGemini 4 Argonnormalised score marginsData Nov 15, 2026Data: Vals AI, Artificial Analysis, Mercor and 45 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…