Atlas

Compare Models

Opus 4.8 vs Gemini 4 Argon

Gemini 4 Argon leads by 6.2 capability points (≈95% interval +4.0 to +8.4): a clear lead over Opus 4.8 · P(higher) ≈ 100%

Wins 141 of 176 shared benchmarks (22 lost, 13 tied)

Capability vs. relative cost
Opus 4.8Gemini 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.

Opus 4.8vsGemini 4 Argoncapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins22 – 141 · 176 shared, 13 tied

Comparing matching setups only (35 benchmarks excluded) · suites counted once
← Opus 4.8Gemini 4 Argon →

Opus 4.8vsGemini 4 Argonwins by categoryData Nov 15, 2026Data: Vals AI, Center for AI Safety, benchmarks.bio 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 Opus 4.8.

ARC-AGI-3 (Semi-Private)General QA1.597.5+95.9
Concept Synth INDUCTION Challenge64 Round 1General QA6.365.6+59.4
HPCT Refusal (CAIS)Safety85.029.0−56.0
HieroglyphBenchMultimodal14.069.7+55.7
LLM Chess EloGames4501609+47.6
EnigmaEvalGeneral QA20.564.6+44.1
BrowseComp Extended (Mercor held-out)Search30.072.0+42.0
Time Horizon Index: KSPAgents11.852.5+40.7
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
Terminal-Bench Science 0.1 (Vals)Science4.344.3+40.0
Blueprint-Bench 2Multimodal0.10.5+39.9
TaxCalcBench TY25 (web search)Finance & Accounting36.074.0+38.0
TextQuests Game ProgressGames40.177.3+37.2
Agent Red Teaming k=100 (CAIS)Safety47.010.2+36.8
Terminal-Bench 4.0Code23.657.4+33.8
MASKSafety94.461.0−33.4
Terminal-Bench 4.0 (Artificial Analysis)Code21.755.1+33.3
AutomationBench 1.0.6 (Zapier)Agents16.950.1+33.2
AutomationBench-AA 1.0.6Agents45.677.5+31.9
Humanity's Sixth SenseMultimodal23.555.4+31.9
MindCube TinyMultimodal64.996.7+31.8
ChessBench EloGames9611485+31.2
VCT-Refusal (CAIS)Safety41.210.7−30.5
TxBench Antibody Discovery V1 (100 evaluations)Science33.363.7+30.3
Furniture Assembly BenchmarkMultimodal42.572.5+30.0
SpatialViz-BenchMultimodal65.294.2+29.0
RiemannBenchMath47.276.0+28.8
BullshitBench v2 LeaderboardSafety94.066.0−28.0
Mystery Game PuzzlesGames36.064.0+28.0
ERQAMultimodal59.586.5+27.0
NoCoT-Bench NCRI 15.2General QA102151+27.0
GameBench 2Games62.589.2+26.7
HiL-Bench (Human-in-Loop Benchmark)Agents35.362.0+26.7
MineBench — Bradley-Terry RatingGames17862195+24.8
Bug Hunt BenchCode14.338.7+24.4
Terminal-Bench 2.1Code82.057.7−24.3
Finance Agent v2 - PrecedentsFinance & Accounting25.649.8+24.2
DeepSWE v1.1Code54.477.9+23.5
BuseyBench SVGMultimodal5.27.6+23.5
APEX-Agents 1.1Professional Work64.487.4+23.0
SimpleQA Verified (Epoch anti-abstention)General QA53.076.0+23.0
ALE-BenchCode21022747+22.9
HardGeoBench <=100miMultimodal36.358.8+22.5
FrontierMath Tier 4 (v2) - PrivateMath56.178.0+22.0
Code MigrationCode47.368.2+20.9
Harvey's Legal Agent Benchmark - Antitrust CompetitionLegal0.020.0+20.0
Harvey's Legal Agent Benchmark - Capital MarketsLegal20.040.0+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 - Emerging Companies Venture CapitalLegal20.040.0+20.0
Harvey's Legal Agent Benchmark - Healthcare Life SciencesLegal40.020.0−20.0
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 - Structured Finance SecuritizationLegal20.00.0−20.0
IntPhys 2Multimodal55.174.6+19.5
Humanity's Last Exam DiamondGeneral QA23.142.5+19.4
ARC-AGI-2General QA72.191.3+19.2
LAB-Bench ProtocolQA (SecureBio)Science76.995.9+19.0
DrugDiscoveryBenchScience46.865.4+18.6
ComplexConstraintsChat & Writing35.653.3+17.7
GDPval Extended (Mercor held-out rubric score)Professional Work58.976.4+17.5
HardGeoBench <=10miMultimodal27.545.0+17.5
TextQuests Average Cumulative HarmSafety21.026.1−17.4
OncologyBench Strict (25 tasks, October 9 rubric)Health & Medicine16.033.3+17.3
Agents' Last Exam (ALE)Agents22.439.5+17.1
LMArena Code Arena WebDevCode15551682+16.6
HardGeoBench Avg DistanceMultimodal1245741+15.5
MedXpertQA MM Extended (Mercor held-out)Health & Medicine51.666.7+15.1
PCT Political Inconsistency (CAIS)Safety53.838.7+15.1
LMArena Agent Arena Bash RecoveryAgents7.2-21.4−14.3
Tax Agent BenchFinance & Accounting64.979.0+14.1
GBENCH One-shot Coding GScoreGames54.067.9+13.8
TxBench Oligo Discovery V1 (113 evaluations)Science37.551.0+13.6
FrontierCode 2.0 - Pass RateCode17.130.6+13.5
BioSecBench-Surveillance V1 Clean RerunsScience46.959.8+12.9
ChessBench CoherenceGames79.792.6+12.9
Finance Agent v2 - Earnings AnalysisFinance & Accounting72.184.8+12.7
Finance Agent v2 - Market AnalysisFinance & Accounting67.279.7+12.5
LisanBench Difficulty-WeightedGames26943849+12.4
FrontierCode 2.0 - ScoreCode15.327.6+12.3
Finance Agent v2 - All PassFinance & Accounting43.555.8+12.3
Design Arena Overall Code CategoriesCode12521326+12.3
Finance Agent v2 - AdjustmentsFinance & Accounting48.460.5+12.1
ScBench-Long V1 Clean RerunsScience16.528.4+11.9
AA-LCR Extended (Mercor held-out)Search63.274.7+11.5
PostTrainBench Average (v1.1)Indexes33.845.3+11.5
VCT v2 (SecureBio)Science45.957.3+11.4
HPCT (SecureBio)Science63.174.2+11.1
Legal Research BenchLegal43.854.8+11.1
Artificial Analysis Intelligence Index v4.3.2Indexes41.852.6+10.8
Finance Agent v2 - General Qualitative AnalysisFinance & Accounting66.477.0+10.5
VCT (SecureBio)Science48.258.4+10.2
ChessBench AccuracyGames72.481.8+10.1
HardGeoBench Country AccuracyMultimodal53.863.8+10.0
Finance Agent v2 - General Quantitative AnalysisFinance & Accounting71.681.5+9.9
AA-Briefcase v1.1 EloAgents13201485+9.0
Finance Agent v2 - Disclosure AnalysisFinance & Accounting65.774.5+8.8
Remote Labor Index (RLI)Professional Work8.317.1+8.7
GPQA-Bio (SecureBio)Science80.989.3+8.4
Artificial Analysis Progress Index v1Indexes318342+8.2
GPQA DiamondGeneral QA88.496.4+8.0
VariantBench V1 Clean RerunsScience42.450.3+7.9
NoCoT-Bench NCKI (kspine v3)General QA117128+7.8
GDPval-AA v2.1Professional Work14611634+7.6
BioTIER RefusalSafety94.787.2−7.5
SciCode 1.0.1 (Artificial Analysis)Code54.461.8+7.4
LiveBench 2.0Agents76.884.1+7.3
Bio Capabilities Index (SecureBio)Indexes164174+7.1
LMArena Text Arena EloChat & Writing14821522+7.1
Chess Puzzles (task version 2.0)Games52.059.0+7.0
TxBench Preclinical Pharmacology V1 Clean RerunsScience57.764.7+7.0
MBCT (SecureBio)Science56.463.2+6.8
AA-Omniscience IndexIndexes28.842.4+6.8
Vals RSI Index v1.1 (four tasks)Agents24.030.6+6.6
PubMedQA (SecureBio)Health & Medicine75.982.4+6.5
Finance Agent v2 - Financial ModelingFinance & Accounting23.329.8+6.5
Harvey HLAB — Criteria Pass RateLegal87.994.0+6.2
LMCAGeneral QA57.563.7+6.2
Finance Agent v2 - ComparablesFinance & Accounting44.951.0+6.1
ARC-AGI-1 Semi-PrivateGeneral QA92.598.5+6.0
MetagenomicsBench V1Science48.342.3−6.0
FrontierMath Tiers 1-3 (v2) - PrivateMath80.086.0+6.0
LisanBench Path LengthGames94637560−5.9
Excel Modeling BenchmarkFinance & Accounting69.475.2+5.9
Humanity's Last ExamGeneral QA49.855.6+5.8
Conceptual Reasoning IndexIndexes68.372.4+5.7
WCB (SecureBio)Science66.761.1−5.6
MedCodeHealth & Medicine53.258.8+5.6
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & Medicine19.413.9−5.6
Wiz Cyber Model Arena v3.9Cybersecurity65.470.8+5.4
Epoch Capabilities Index (ECI)Indexes160167+5.4
APEX-SWECode43.949.2+5.3
Arena Alignment IndexSafety74.579.4+4.9
SpatialBench-Long V1 Clean RerunsScience33.328.6−4.8
LegalBenchLegal83.688.3+4.7
GDP.pdfProfessional Work24.028.4+4.4
LAB-Bench CloningScenarios (SecureBio)Science95.9100.0+4.1
HealthBench ProfessionalHealth & Medicine60.364.2+3.9
Machiavelli Deception (CAIS)Safety83.785.0−3.8
AA-AnalystAgent (pass^5)Professional Work45.048.8+3.8
SpatialBench V1 Clean RerunsScience65.869.0+3.2
EpiBench V1 Clean RerunsScience25.828.9+3.1
SpeciEval (Pooled English Runs)Safety84.381.3−3.1
Senior SWE-BenchCode30.527.7−2.8
DistressBenchSafety71.468.6−2.8
DTBenchGeneral QA94.997.6+2.7
ScBench V1 Clean RerunsScience57.159.1+2.1
MedScribeHealth & Medicine85.887.4+1.7
OTIS Mock AIME 2024-2025Math98.3100.0+1.7
Public Benefits Bench v1.1Professional Work68.169.8+1.6
ProgramBenchCode1.02.5+1.5
GDPval (Mercor 213-task rubric score)Professional Work84.585.9+1.4
BioSecBench-Function V1 Clean RerunsScience43.144.4+1.4
HealthBenchHealth & Medicine58.860.0+1.2
MMLU-Bio (SecureBio)Science90.992.1+1.2
SAGEMath54.853.6−1.1
FrontierCode 1.1 Extended - ScoreCode59.660.6+1.0
WMDP-Bio (SecureBio)Science92.391.5−0.8
FrontierCode 1.1 Extended - Pass RateCode65.566.1+0.6
LLM Creative Story WritingChat & Writing0.80.8+0.5
ACCoRDGeneral QA77.477.3−0.1
ABC-Bench liquid handling (SecureBio)Science100.0100.00.0
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 - 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 - 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 - TaxLegal0.00.00.0
Harvey's Legal Agent Benchmark - White Collar Defense InvestigationsLegal0.00.00.0

Opus 4.8vsGemini 4 Argonnormalised score marginsData Nov 15, 2026Data: Vals AI, Center for AI Safety, benchmarks.bio 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…