Atlas

Compare Models

Gemini 4 Argon vs Grok 4.7

Gemini 4 Argon leads by 5.3 capability points (≈95% interval +3.0 to +7.6): a clear lead over Grok 4.7 · P(higher) ≈ 100%

Wins 208 of 274 shared benchmarks (49 lost, 17 tied)

Capability vs. relative cost
Gemini 4 ArgonGrok 4.7Closed 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.

Gemini 4 ArgonvsGrok 4.7capability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins208 – 49 · 274 shared, 17 tied

Comparing matching setups only (44 benchmarks excluded) · suites counted once
← Gemini 4 ArgonGrok 4.7 →

Gemini 4 ArgonvsGrok 4.7wins by categoryData Nov 15, 2026Data: Vals AI, Artificial Analysis, Surge AI and 70 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Grok 4.7 minus Gemini 4 Argon.

ARC-AGI-3 (Semi-Private)General QA97.310.0−87.3
ProofBench v1.1Math99.026.0−73.0
CheatBenchSafety8.678.0−69.4
FrontierMath Tier 4 (v2) - PrivateMath78.017.1−61.0
AA-PersonalityBench Slop indexChat & Writing67.68.6−59.6
ChartographyMultimodal71.614.7−56.9
Furniture Assembly BenchmarkMultimodal72.520.8−51.7
AA-PersonalityBench Informativeness indexChat & Writing82.933.4−50.0
LAB-Bench ProtocolQA (SecureBio)Science95.951.4−44.5
ArXivMath 09/2026 (Native Harnesses)Math87.745.3−42.5
IOI (Vals v2)Code100.057.7−42.3
HardGeoBench Avg DistanceMultimodal7412104−42.0
Toloka Arena (v3)Agents73.031.0−42.0
TreasuryQA BenchProfessional Work60.018.1−41.9
WARC-Bench (192-task test split)Agents76.635.6−41.0
SRE Bench — Fully SolvedCode44.33.8−40.5
SAFE-Teen Health Advice (no prompt)Safety56.216.1+40.1
Harvey's Legal Agent Benchmark - Emerging Companies Venture CapitalLegal40.00.0−40.0
Harvey's Legal Agent Benchmark - Trusts Estates Private ClientLegal40.00.0−40.0
SAFE-Teen 106: Health Advice (no prompt)Safety57.418.7+38.7
Terminal-Bench-Science 0.1Science48.611.4−37.1
AA-PersonalityBench Transparency indexChat & Writing21.958.1+36.5
ArXivMath 08/2026 (Native Harnesses)Math82.546.5−36.0
Mystery Game PuzzlesGames64.029.0−35.0
ALE-BenchCode28051823−34.8
AA-PersonalityBench Warmth indexChat & Writing79.144.8−34.6
Terminal-Bench Science 0.1 (Vals)Science44.310.0−34.3
Concept Synth INDUCTION Challenge100 Round 1General QA57.024.0−33.0
FrontierMath Tiers 1-3 (v2) - PrivateMath86.053.0−33.0
BrowseComp Extended (Mercor held-out)Search72.039.5−32.5
LLM Chess EloGames1609822−32.3
SAFE-Teen 106 worst branch: Health Advice (no prompt)Safety70.839.2+31.7
Concept Synth INDUCTION Challenge64 Round 1General QA65.634.4−31.2
HardGeoBench <=100miMultimodal58.827.5−31.2
BuseyBench SVGMultimodal7.64.6−29.9
ARC-AGI-2General QA91.361.4−29.9
Terminal-Bench 4.0 (Artificial Analysis)Code55.125.8−29.3
VIALSMultimodal47.218.2−29.0
FrontierML ($200 wallet)Agents39.510.9−28.6
FrontierSWE V2 Worst@5Code44.515.9−28.6
TaxCalcBench TY25 (web search)Finance & Accounting74.046.0−28.0
HardGeoBench <=10miMultimodal45.017.5−27.5
Terminal-Bench 2.1Code57.785.0+27.3
AA-PersonalityBench Organization indexChat & Writing56.229.6−26.9
FrontierMemory v1.1Agents64.437.8−26.7
τ³-Banking (Mercor run, bm25_grep retrieval)Agents58.132.6−25.5
SAFE-Teen 106 worst branch: Health Advice (under-18 prompt)Safety43.619.2+24.4
ChessBench EloGames14851076−24.3
Finance Agent v2 - PrecedentsFinance & Accounting49.825.9−23.9
Bio Capabilities Index (SecureBio)Indexes174141−23.5
Code MigrationCode68.244.8−23.4
FrontierSWE V2 Best@5Code64.341.1−23.2
AA-3D v1.0 EloCode1147992−22.4
Visa Agentic Commerce BenchmarkAgents82.260.0−22.2
BrokenArXiv 09/2026 (Native Harnesses)Math49.727.7−22.0
Blueprint-Bench 2Multimodal0.50.3−21.9
DataCenterBench Tier 1Agents35.013.3−21.7
HoneyBench v0.2 (hack rate)Safety63.685.1−21.5
MineBench — Bradley-Terry RatingGames21951848−21.0
Vals Science IndexIndexes55.434.8−20.5
HPCT (SecureBio)Science74.253.7−20.5
MysteryMechanismScience45.525.2−20.3
Harvey's Legal Agent Benchmark - Corporate GovernanceLegal20.00.0−20.0
Harvey's Legal Agent Benchmark - Data Privacy CybersecurityLegal20.00.0−20.0
Harvey's Legal Agent Benchmark - Funds Asset ManagementLegal40.020.0−20.0
Harvey's Legal Agent Benchmark - ImmigrationLegal20.00.0−20.0
Harvey's Legal Agent Benchmark - International Trade SanctionsLegal20.00.0−20.0
Harvey's Legal Agent Benchmark - Litigation Dispute ResolutionLegal0.020.0+20.0
HardGeoBench Country AccuracyMultimodal63.843.8−20.0
SimpleQA Verified (Epoch anti-abstention)General QA76.056.0−20.0
LAB-Bench CloningScenarios (SecureBio)Science100.080.1−19.9
Terminal-Bench 4.0Code57.437.6−19.8
AA-PersonalityBench Conviction indexChat & Writing63.882.9+19.2
BrokenArXiv 08/2026 (Native Harnesses)Math54.235.1−19.1
SAFE-Teen Health Advice (under-18 prompt)Safety29.610.6+19.0
Humanity's Last Exam DiamondGeneral QA42.523.6−18.9
SAFE-Teen 106: Health Advice (under-18 prompt)Safety30.812.0+18.8
APEX-Agents 1.1Professional Work87.468.7−18.7
Chess Puzzles (task version 2.0)Games59.041.0−18.0
SAFE-Teen 106: Urgent Safety (no prompt)Safety33.115.1+18.0
Finance Agent v2 - Market AnalysisFinance & Accounting79.761.8−17.9
SAFE-Teen 106 worst branch: Relationships (no prompt)Safety65.747.9+17.8
DAYJOB: LawLegal9.927.6+17.7
SAFE-Teen Urgent Safety (no prompt)Safety33.315.7+17.6
Code Migration v1.1Code70.853.2−17.6
CritPt v2 (agentic, avg@5)Science72.955.7−17.1
MBCT (SecureBio)Science63.246.1−17.1
OpenWorldEngineer (Mean@8)Code24.17.0−17.0
Ramp Accounting BenchFinance & Accounting61.144.1−17.0
GameBench 2Games89.272.3−16.8
RiemannBench (November 2026 rescoring)Math74.457.6−16.8
DAYJOB: Data ScienceProfessional Work23.36.7−16.6
DistressBenchSafety68.652.4−16.2
APEX-Accounting 1.1Finance & Accounting71.555.3−16.2
GDP.xlsxProfessional Work38.322.3−16.0
LMArena Agent Arena Net Improvement (Steering Burden)Agents10.72.6−15.5
VCT v2 (SecureBio)Science57.342.0−15.3
SpeciEval (Pooled English Runs)Safety81.396.4+15.1
Artificial Analysis Intelligence Index v5.0.1Indexes50.435.5−14.9
MedXpertQA MM Extended (Mercor held-out)Health & Medicine66.752.0−14.7
OncologyBench Strict (25 tasks, October 9 rubric)Health & Medicine33.318.7−14.7
LMCAGeneral QA63.749.4−14.3
LMArena Agent Arena Bash RecoveryAgents-21.47.2+14.3
ATLAS (Exa, 200-task subset)Search53.839.5−14.3
Finance Agent v2 - AdjustmentsFinance & Accounting60.546.3−14.3
LMArena Text Arena EloChat & Writing15221442−14.0
Humanity's Last Exam (Text-Only)General QA57.143.1−13.9
Finance Agent v2 - All PassFinance & Accounting55.842.1−13.7
Westworld Finance Diligence BenchFinance & Accounting60.747.2−13.6
SnorkelLegalLegal46.032.8−13.2
SWE-bench Multimodal (Vals v1, 480 Tasks)Code56.042.9−13.1
Finance Agent v2 - General Quantitative AnalysisFinance & Accounting81.568.6−13.0
Creative Writing v3 EloChat & Writing17582012+12.9
SAGEMath53.640.8−12.9
FrontierBioMLScience36.824.0−12.9
Physical-Engineering's Last Exam (hidden test)Professional Work40.627.8−12.8
Conceptual Reasoning IndexIndexes72.463.1−12.8
AgenticRL-Bench (Verifiable)Agents14.01.8−12.3
Finance Agent v2 - Financial ModelingFinance & Accounting29.817.7−12.1
RWE-benchAgents12.00.0−12.0
AutomationBench-AA 1.0.6Agents77.565.6−12.0
GBENCH One-shot Coding GScoreGames67.956.3−11.6
AA-PersonalityBench Imagination indexChat & Writing46.735.3−11.5
Dual-Arm Sim (task completion)Agents11.30.0−11.2
Real-SWE 1.1Code39.228.0−11.2
MKA (Artificial Analysis)Professional Work51.340.1−11.1
Bug Hunt BenchCode38.727.6−11.1
RSI-Exam 0.1 (Full 88)Agents0.50.4−11.0
Agent Initiative BenchmarkAgents5.64.5−10.9
Code2Video-AA EloCode12591127−10.8
LMArena Agent Arena Praise vs ComplaintAgents24.13.1−10.5
LegalReasonBenchLegal50.840.4−10.4
Vals RSI Index v1.1 (four tasks)Agents30.620.2−10.4
CWE-bench v2 Pass@1Cybersecurity65.054.7−10.3
Finance Agent v2 - Earnings AnalysisFinance & Accounting84.874.6−10.2
Snorkel SWE-bench CLICode16.36.1−10.2
MMMU ProMultimodal88.378.2−10.2
Tax Agent BenchFinance & Accounting79.069.0−10.0
Finance Agent v2 - Disclosure AnalysisFinance & Accounting74.564.5−10.0
LiveBench 2.0Agents84.174.1−10.0
Snorkel Agentic Coding 2.0Code41.831.9−9.9
AA-PersonalityBench Expressiveness indexChat & Writing84.875.3−9.6
Design Arena Overall Code CategoriesCode13261270−9.3
MedCodeHealth & Medicine58.849.6−9.2
CritPt (avg@5)Science27.118.0−9.1
Snorkel Digital Work IndexIndexes28.319.3−9.0
BioResearchBenchScience13.04.0−9.0
SAFE-Teen 106 worst branch: Relationships (under-18 prompt)Safety14.05.1+8.9
HLAB-Vals-1.1Legal27.118.3−8.7
Finance Agent v2 - ComparablesFinance & Accounting51.042.2−8.7
Epoch Capabilities Index (ECI)Indexes167156−8.5
CyberBench v1.1 OverallCybersecurity77.969.5−8.4
ChessBench AccuracyGames81.874.0−8.4
ARC-AGI-1 Semi-PrivateGeneral QA98.590.2−8.3
PLE (Physical-Engineering's Last Exam)Professional Work40.231.9−8.3
Excel Modeling BenchmarkFinance & Accounting75.267.0−8.3
Finance Agent v2 - General Qualitative AnalysisFinance & Accounting77.068.7−8.2
SAFE-Teen 106 worst branch: Urgent Safety (no prompt)Safety52.644.4+8.2
SWE Atlas v1.1 - OverallCode70.162.1−8.0
AA-Briefcase v1.1 EloAgents14851629+7.9
BenchBench-CloningScience18.810.9−7.9
Harvey LAB (Mercor run, undisclosed ~213-task cohort)Legal27.419.5−7.9
LMArena Agent Arena Steering BurdenAgents13.2-2.3−7.7
Legal Research BenchLegal54.847.1−7.7
SnorkelManufacturingProfessional Work19.512.1−7.4
HANDBOOK.md Agents v1.1Professional Work36.644.0+7.4
LMArena Agent Arena Bash ReliabilityAgents-5.49.3+7.3
DeepFinanceFinance & Accounting41.434.1−7.3
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & Medicine13.921.1+7.2
SnorkelUnderwrite 2.0Finance & Accounting23.330.5+7.2
Vals Finance IndexIndexes66.058.8−7.2
AutoResearchExam (hidden-test AUARC)Agents0.60.5−7.2
SnorkelRevOpsProfessional Work18.411.3−7.1
Finance Agent v2.1Finance & Accounting67.760.6−7.1
BioMysteryBench (Vals Terminus 2)Science76.369.3−7.0
Eval Connections One-Shot with TrapsGames92.085.0−7.0
HealthBench (length-adjusted, Artificial Analysis)Health & Medicine49.156.0+7.0
DeepSWE v1.1Code77.971.0−6.9
GPQA-Bio (SecureBio)Science89.382.5−6.8
Artificial Analysis Progress Index v1Indexes342323−6.7
Artificial Analysis Coding Agent Index v2.0Indexes53.246.6−6.6
CursorBench 4.0Code52.946.3−6.6
AA-LCR Extended (Mercor held-out)Search74.768.2−6.5
ChessBench CoherenceGames92.686.3−6.3
AA-AnalystAgent (pass^5)Professional Work48.855.0+6.3
Artificial Analysis Intelligence Index v4.3.2Indexes52.646.4−6.1
BioSimBench 1.0Science18.412.3−6.1
Snorkel WorkplaceAgentsProfessional Work21.215.2−6.0
HealthBenchHealth & Medicine60.054.2−5.9
GDP.pdfProfessional Work28.422.8−5.6
LMArena Code Arena WebDevCode16821641−5.4
Access to Justice BenchmarkLegal68.573.7+5.2
AA-Omniscience IndexIndexes42.432.0−5.2
SnorkelFinance 2.0Finance & Accounting15.120.2+5.1
o11y-bench 1.1 Pass^3Agents81.376.6−4.7
MissionReportProfessional Work64.159.5−4.6
ComplexConstraintsChat & Writing53.348.9−4.4
DAYJOB: FinanceFinance & Accounting18.814.5−4.3
DAXBench v2Code88.392.6+4.3
APEX-SWECode49.253.4+4.2
LLM Creative Story WritingChat & Writing0.80.4−4.2
Public Benefits Bench v1.1Professional Work69.865.6−4.1
Vals Legal IndexIndexes76.872.7−4.1
FrontierPhysics v0.1Science8.34.2−4.1
LMArena Agent Arena Confirmed SuccessAgents13.25.1−4.1
SciCode 1.0.1 (Artificial Analysis)Code61.857.8−4.1
Next.js AI Agent EvaluationsAgents100.096.0−4.0
LegalBenchLegal88.384.4−3.9
SAFE-Teen Relationships (no prompt)Safety28.632.5−3.9
Artificial Analysis Coding Agent Index v2.1Indexes45.541.6−3.9
GDPval-AA v2.1Professional Work16341718+3.7
GPQA DiamondGeneral QA96.492.7−3.7
Harvey HLAB — Criteria Pass RateLegal94.090.5−3.5
SAFE-Teen 106 worst branch: Urgent Safety (under-18 prompt)Safety30.226.7+3.5
WCB (SecureBio)Science61.157.6−3.5
GDPxProfessional Work7.33.9−3.5
Excel Modeling Benchmark v1.1Finance & Accounting74.871.5−3.3
sudo L7Code18.915.6−3.3
Arena Alignment IndexSafety79.482.7+3.3
ACCoRDGeneral QA77.374.0−3.3
PubMedQA (SecureBio)Health & Medicine82.479.2−3.2
WitnessBench v2.0Games17.914.8−3.1
SAFE-Teen 106: Relationships (no prompt)Safety28.531.5−3.0
Political Science LLM BenchmarkProfessional Work69.666.6−3.0
DualEntry Accounting AI BenchmarkFinance & Accounting91.188.1−3.0
WMDP-Bio (SecureBio)Science91.588.8−2.7
DAYJOB: HealthcareHealth & Medicine10.88.4−2.4
SAFE-Teen Urgent Safety (under-18 prompt)Safety12.310.1+2.2
Longform Creative WritingChat & Writing80.983.0+2.1
SAFE-Teen 106: Urgent Safety (under-18 prompt)Safety12.110.1+2.0
ABC-Bench liquid handling (SecureBio)Science100.098.0−2.0
DAYJOB: AccountingFinance & Accounting18.020.0+2.0
ProgramBenchCode2.50.5−2.0
WellnessBenchSafety50.048.0−2.0
SAFE-Teen 106: Relationships (under-18 prompt)Safety3.01.0+2.0
Creative Writing v3 Rubric ScoreChat & Writing15.615.9+2.0
MedScribeHealth & Medicine87.489.4+1.9
OTIS Mock AIME 2024-2025Math100.098.1−1.9
SAFE-Teen Relationships (under-18 prompt)Safety3.01.1+1.9
SWE-Marathon v1.1Code43.845.6+1.9
DAYJOB: TaxFinance & Accounting3.25.0+1.8
AML BenchFinance & Accounting64.165.8+1.7
FrontierCode 2.0 - Pass RateCode30.632.3+1.7
DTBenchGeneral QA97.696.0−1.6
FrontierMath ErdősMath1.50.0−1.5
FrontierCode 2.0 - ScoreCode27.628.9+1.3
FrontierCode 1.1 Extended - ScoreCode60.659.4−1.2
MMLU-Bio (SecureBio)Science92.190.9−1.2
CloudOpsBench v1.0Agents57.556.5−1.0
Financial Coaching BenchFinance & Accounting73.174.1+0.9
FrontierCode 1.1 Extended - Pass RateCode66.165.2−0.9
SAGE v1.1Math89.588.7−0.8
Robo-benchAgents0.80.2−0.6
AA-LCR v1.1Search79.780.0+0.3
Senior SWE-BenchCode27.727.4−0.3
LMArena Agent Arena Tool HallucinationAgents0.51.0+0.2
Immigration BenchLegal68.168.00.0
CWE-bench v1 Programmatic Pass@1Cybersecurity68.068.00.0
Harvey's Legal Agent Benchmark - Antitrust CompetitionLegal20.020.00.0
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 - Capital MarketsLegal40.040.00.0
Harvey's Legal Agent Benchmark - Corporate M&ALegal20.020.00.0
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 - Intellectual PropertyLegal20.020.00.0
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

Gemini 4 ArgonvsGrok 4.7normalised score marginsData Nov 15, 2026Data: Vals AI, Artificial Analysis, Surge AI and 70 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…