Atlas

Compare Models

Claude Opus 5.5 vs GPT-6 Sol

Claude Opus 5.5 leads by 4.0 capability points (≈95% interval +1.5 to +6.5): a clear lead over GPT-6 Sol · P(higher) ≈ 100%

Wins 169 of 244 shared benchmarks (48 lost, 27 tied)

Capability vs. relative cost
Claude Opus 5.5GPT-6 SolClosed 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 Opus 5.5vsGPT-6 Solcapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins169 – 48 · 244 shared, 27 tied

Comparing matching setups only (129 benchmarks excluded) · suites counted once
← Claude Opus 5.5GPT-6 Sol →

Claude Opus 5.5vsGPT-6 Solwins by categoryData Nov 15, 2026Data: Mercor, Vals AI, Artificial Analysis and 65 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is GPT-6 Sol minus Claude Opus 5.5.

KernelBench Mega Normalized (RTX PRO 6000)Code100.02.7−97.3
Vending-Bench Arena: Round 13Games809410530+93.7
AA-PersonalityBench Informativeness indexChat & Writing90.510.5−80.8
AA-PersonalityBench Organization indexChat & Writing79.112.4−67.3
AA-PersonalityBench Imagination indexChat & Writing94.329.6−65.4
AA-PersonalityBench Expressiveness indexChat & Writing90.533.4−57.7
AA-PersonalityBench Transparency indexChat & Writing90.539.1−51.9
Medical Long Context Reasoning (MLCR-AA)Health & Medicine66.716.1−50.6
YC-BenchGames28737251441793−50.0
LisanBench Path LengthGames240298222−49.1
Terminal-Bench-Science 0.1Science61.917.6−44.3
TxBench Oligo Discovery V1 Clean RerunsScience1.440.6+39.2
Olam Social Poker EloGames15061542+39.1
TxBench Preclinical Pharmacology V1 Clean RerunsScience26.363.0+36.7
BioSecBench-Function V1 Clean RerunsScience73.639.3−34.3
KernelBench CUDA Normalized (RTX PRO 6000)Code76.342.0−34.2
BioSecBench-Refusal V1 Clean RerunsSafety3.837.9+34.1
EnterpriseBench: CoreCraftProfessional Work74.443.1−31.3
Search ArenaSearch12131174−31.2
AA-PersonalityBench Warmth indexChat & Writing84.854.3−30.8
LisanBench Difficulty-WeightedGames59633118−30.5
TxBench Antibody Discovery V1 Clean RerunsScience6.837.0+30.2
Terminal-Bench 4.0 (Artificial Analysis)Code59.630.3−29.3
SWE-bench Verified Extended (Mercor held-out)Code66.039.3−26.7
HardGeoBench Avg DistanceMultimodal6691497−25.6
Furniture Assembly BenchmarkMultimodal83.358.3−25.0
LMArena Agent Arena Net Improvement (Steering Burden)Agents19.56.9−23.8
HardGeoBench <=100miMultimodal61.337.5−23.8
Decision-Bench v1.1Professional Work59.136.5−22.6
APEX-SWECode67.645.0−22.6
ScBench-Long V1 Clean RerunsScience6.828.4+21.6
HardGeoBench Country AccuracyMultimodal67.546.3−21.2
DAYJOB: HealthcareHealth & Medicine24.73.6−21.1
Terminal-Bench 4.0Code64.843.8−21.0
Harvey's Legal Agent Benchmark - Energy Natural ResourcesLegal20.00.0−20.0
Harvey's Legal Agent Benchmark - Healthcare Life SciencesLegal20.00.0−20.0
LAB-Bench LitQA2 (SecureBio)Science71.151.8−19.3
AA-PersonalityBench Conviction indexChat & Writing69.550.5−19.2
AA-PersonalityBench Slop indexChat & Writing20.01.0−19.2
EBR-bench v4 (card ban, single agent)Games71.453.3−18.1
SWE-bench VerifiedCode98.280.3−17.9
AA-Briefcase v1.1 EloAgents18011476−17.8
ProgramBench (Mercor 100-task subset)Code20.02.3−17.7
Terminal-Bench Science 0.1 (Vals)Science48.631.4−17.1
GDPval Extended (Mercor held-out rubric score)Professional Work75.858.7−17.1
Toloka Arena (v3)Agents75.058.0−17.0
ProofBench v1.1Math100.083.0−17.0
EpiBench V1 Clean RerunsScience39.022.0−17.0
HardGeoBench <=10miMultimodal42.526.3−16.2
HealthBenchHealth & Medicine64.348.2−16.0
FrontierCUA pass@1Agents36.020.0−16.0
SM BenchSafety81.866.0−15.8
VoxelBench Text RatingGames26132394−15.6
GDPval-AA v2.1Professional Work18661514−15.5
BioSecBench-Structure V1 (58 evaluations)Science29.313.8−15.5
GameBench 2Games94.879.4−15.3
SimpleBenchGeneral QA88.473.1−15.3
LMArena Agent Arena Praise vs ComplaintAgents40.310.0−15.2
Design Arena Overall Code CategoriesCode13841293−15.1
Mystery Game PuzzlesGames71.056.0−15.0
SciCode (main problems)Science27.312.7−14.6
DAYJOB: FinanceFinance & Accounting23.99.3−14.6
LegalReasonBenchLegal46.531.9−14.6
Blueprint-Bench 2Multimodal0.50.4−14.3
DeepsecBench — Revalidated V1Code26.740.9+14.2
DeepsecBench — Revalidated V1 (233 issues)Code26.640.7+14.2
TaxCalcBench TY25Finance & Accounting42.028.0−14.0
SWE-bench MultilingualCode95.581.7−13.8
Financial Coaching BenchFinance & Accounting76.262.6−13.5
Humanity's Sixth SenseMultimodal44.631.2−13.4
GDP.xlsxProfessional Work30.317.1−13.2
LAB-Bench ProtocolQA (SecureBio)Science94.081.1−12.9
APEX-Agents 1.1Professional Work81.368.7−12.6
τ³-Banking (Mercor run, bm25_grep retrieval)Agents55.043.0−12.0
SWE Atlas Codebase QnA Extended (Mercor held-out)Code43.331.3−12.0
VariantBench V1 Clean RerunsScience50.638.7−11.9
SimpleQA Verified (Epoch anti-abstention)General QA72.260.7−11.5
HieroglyphBenchMultimodal37.426.0−11.4
CVDP (517-problem set)Code64.152.7−11.4
SpatialBench-Long V1 Clean RerunsScience20.231.6+11.3
SpeciEval (Pooled English Runs)Safety87.298.5+11.3
SpeciEval (Refusal-Aware Retries, Historical)Safety87.298.5+11.3
Humanity's Last Exam (Text-Only)General QA57.546.3−11.2
ITBench-AAAgents38.249.4+11.2
VCT v2 (SecureBio)Science60.549.3−11.2
Chess Puzzles (task version 2.0)Games67.056.0−11.0
FrontierCode 2.0 (Artificial Analysis)Code46.435.6−10.9
LibraryDesignBenchCode48.938.5−10.4
DeepResearch Bench II (Mercor 130-task set)Search53.342.9−10.4
AA-3D v1.0 EloCode12021131−10.3
HealthBench ProfessionalHealth & Medicine71.461.2−10.2
AA-Omniscience IndexIndexes46.427.1−9.6
Finance Agent v2Finance & Accounting58.649.0−9.5
Humanity's Last Exam DiamondGeneral QA54.645.2−9.4
MedScribeHealth & Medicine91.482.0−9.4
Conceptual Reasoning IndexIndexes77.971.0−9.4
OncologyBench Strict (25 tasks, October 9 rubric)Health & Medicine29.320.0−9.3
SciCode 1.0.1 (Artificial Analysis)Code66.957.6−9.3
LMCAGeneral QA68.259.1−9.2
Gray Swan IPI Attack Success (k=15)Safety1.010.1−9.1
τ³-Banking (Fireworks, banking_knowledge)Agents56.047.0−9.0
BullshitBench v2 LeaderboardSafety63.054.0−9.0
Professional Reasoning Benchmark - LegalLegal47.638.6−9.0
LMArena Text Arena EloChat & Writing15061457−8.6
Clipping Bench v1.0Games20181240−8.4
ArXivMath 05/2026Math81.790.0+8.3
SWE-TogetherCode68.860.6−8.2
ChessBench CoherenceGames80.188.2+8.1
LMArena Agent Arena Confirmed SuccessAgents20.74.5−8.1
BioSecBench-Surveillance V1 Clean RerunsScience55.847.7−8.1
BrokenArXiv 05/2026Math95.087.0−8.0
ZeroBench Main Questions pass^5Multimodal25.017.0−8.0
Qodo AI Code Review Benchmark (125 PRs)Code56.848.9−7.9
Harvey HLAB — Criteria Pass RateLegal75.883.5+7.7
ArXivMath 06/2026Math83.391.0+7.6
DistressBenchSafety75.668.0−7.6
MBCT (SecureBio)Science68.761.1−7.6
VCT (SecureBio)Science61.454.0−7.4
FACTS Benchmark SuiteGeneral QA65.372.5+7.2
ArXivMath 08/2026 (Native Harnesses)Math86.093.0+7.0
PubMedQA (SecureBio)Health & Medicine87.980.9−7.0
Quant-Arena (iterative total score)Finance & Accounting44.351.1+6.9
BuseyBench SVGMultimodal6.87.5+6.8
SUPERChemScience79.672.8−6.8
Bio Capabilities Index (SecureBio)Indexes171161−6.7
Harvey LAB (Mercor run, undisclosed ~213-task cohort)Legal10.84.1−6.7
ZeroBench Main Mean pass@1 (five samples)Multimodal35.028.8−6.2
ScBench V1 Clean RerunsScience55.461.5+6.2
WMDP-Bio (SecureBio)Science96.190.0−6.1
ACCoRDGeneral QA88.482.4−6.0
DualEntry Accounting AI BenchmarkFinance & Accounting82.276.2−6.0
WCB (SecureBio)Science54.560.5+6.0
APEX-Accounting 1.1Finance & Accounting65.559.7−5.8
DeepSWE v1.1Code72.366.6−5.7
Harvey LAB held-out set (Fireworks run)Legal6.30.7−5.6
SpatialBench V1 Clean RerunsScience72.867.5−5.2
Terminal-Bench 2.1 Extended (Mercor held-out)Code47.542.4−5.1
Roboflow Vision Evals (October 2026, five tasks)Multimodal82.076.9−5.1
FrontierMath Tier 4 (v2) - PrivateMath95.090.0−5.0
GDPval (Mercor 213-task rubric score)Professional Work88.383.5−4.8
RiemannBenchMath69.664.8−4.8
SWE Atlas - Codebase QnACode48.944.1−4.8
Nimble Web ContextSearch71.876.5+4.7
BrokenArXiv 06/2026Math100.095.4−4.6
LiveBench 2.0Agents86.581.9−4.6
MMLU ProGeneral QA93.488.9−4.5
Arena Alignment IndexSafety83.287.6+4.4
Excel Modeling BenchmarkFinance & Accounting75.971.5−4.4
MedXpertQA MM (200-question set)Health & Medicine85.781.3−4.4
LMArena Agent Arena Steering BurdenAgents18.610.1−4.2
MMMU ProMultimodal86.682.5−4.0
AA-LCR Extended (Mercor held-out)Search74.270.2−4.0
GeneBench-Pro Public Release SubsetScience30.026.0−4.0
LMArena Agent Arena Bash RecoveryAgents15.07.0−4.0
ALE-BenchCode27482637−4.0
GBENCH One-shot Coding GScoreGames73.970.1−3.8
Creative Writing v3 EloChat & Writing20512126+3.8
Epoch Capabilities Index (ECI)Indexes169164−3.8
ARC-AGI-2General QA93.389.6−3.7
Epiq LAER CorporateBenchProfessional Work75.479.0+3.6
MacroscopeBenchCode83.980.3−3.6
Vectara Hallucination LeaderboardSafety9.96.5+3.4
SpireClimbGames5.89.2+3.3
ChessBenchGames65.362.0−3.3
Hermes BenchAgents76.773.5−3.2
MedXpertQA MM Extended (Mercor held-out)Health & Medicine71.268.0−3.2
Roboflow Vision Evals (September 2026 protocol)Multimodal85.582.3−3.2
WellnessBenchSafety53.056.0+3.0
ARC-AGI-1 Semi-PrivateGeneral QA98.595.5−3.0
Eval Connections One-Shot with TrapsGames86.089.0+3.0
HealthBench (length-adjusted, Artificial Analysis)Health & Medicine57.054.2−2.8
dfbench Detection RecallCybersecurity54.957.7+2.8
ChessBench EloGames13491303−2.7
MedCodeHealth & Medicine49.847.1−2.7
BrowseComp Extended (Mercor held-out)Search60.563.2+2.7
MetagenomicsBench V0Science39.342.0+2.7
MetagenomicsBench V1Science39.342.0+2.7
GDP.pdfProfessional Work30.628.0−2.6
APEX-1 (Mercor, mean score)Professional Work65.863.6−2.2
LMArena Agent Arena Bash ReliabilityAgents15.110.9−2.1
ChessBench AccuracyGames79.377.4−2.0
GPQA-Bio (SecureBio)Science90.988.9−2.0
ZeroBench (Main questions, pass@5)Multimodal43.041.0−2.0
MMLU-Bio (SecureBio)Science92.790.8−1.9
Terminal-Bench 2.1Code83.585.4+1.9
AdvancedIFChat & Writing82.880.9−1.9
Longform Creative WritingChat & Writing84.682.8−1.8
SAB ResearchFinance & Accounting97.295.6−1.7
Harvey LAB: In-House Contracting (held-out)Legal3.01.3−1.7
DTBenchGeneral QA98.997.3−1.6
Extended NYT Connections 940 (Quadratic v1)Games88.590.1+1.6
MentalHealthBenchSafety52.453.9+1.5
Ruby LLM Benchmarks — Program FixerCode77.579.0+1.5
FrontierMath ErdősMath2.91.5−1.5
FrontierMath Tiers 1-3 (v2) - PrivateMath91.289.8−1.4
CharXiv (descriptive and reasoning combined)Multimodal96.495.0−1.4
GlobeBenchGeneral QA26.325.0−1.3
LLM Chess EloGames15751545−1.2
BrokenArXiv 08/2026 (Native Harnesses)Math77.478.6+1.2
BigFinanceBench private holdout (final-answer accuracy)Finance & Accounting48.247.1−1.1
SAGEMath45.844.8−1.0
AA-LCR v1.1Search84.783.7−1.0
BrowseComp (Mercor 130-question subset)Search88.589.4+0.9
CharXiv Extended (Mercor held-out)Multimodal84.783.8−0.9
ComplexConstraintsChat & Writing49.950.8+0.9
MMMU-Pro Extended (Mercor held-out)Multimodal51.051.9+0.9
BioSimBench 1.0Science14.513.6−0.9
CritPt v2 (agentic, avg@5)Science81.780.9−0.9
CritPt (avg@5)Science31.730.9−0.9
ObviousBenchGeneral QA100.099.3−0.7
Harvey LAB-AA v1.1 (hallucination-gated)Legal4.23.6−0.6
Short Form Bench v1.0Games23892394+0.5
ForecastBench Preliminary Dataset Brier IndexForecasting62.262.6+0.4
Creative Writing v3 Rubric ScoreChat & Writing16.016.1+0.3
FORTRESS Average Risk Score (ARS)Safety13.213.5−0.3
SAB RankingForecasting-0.1-0.1+0.2
LMArena Agent Arena Tool HallucinationAgents0.71.0+0.2
AIME 2025Math100.0100.00.0
CadQueryEval (Corrected Geometry)Code100.0100.00.0
Harvey's Legal Agent Benchmark - Antitrust CompetitionLegal0.00.00.0
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 MarketsLegal0.00.00.0
Harvey's Legal Agent Benchmark - Corporate GovernanceLegal0.00.00.0
Harvey's Legal Agent Benchmark - Corporate M&ALegal0.00.00.0
Harvey's Legal Agent Benchmark - Data Privacy CybersecurityLegal0.00.00.0
Harvey's Legal Agent Benchmark - Emerging Companies Venture CapitalLegal0.00.00.0
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 - Funds Asset ManagementLegal0.00.00.0
Harvey's Legal Agent Benchmark - ImmigrationLegal0.00.00.0
Harvey's Legal Agent Benchmark - InsuranceLegal0.00.00.0
Harvey's Legal Agent Benchmark - Intellectual PropertyLegal0.00.00.0
Harvey's Legal Agent Benchmark - International Trade SanctionsLegal0.00.00.0
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 - Trusts Estates Private ClientLegal0.00.00.0
Harvey's Legal Agent Benchmark - White Collar Defense InvestigationsLegal0.00.00.0
Next.js AI Agent EvaluationsAgents100.0100.00.0
OTIS Mock AIME 2024-2025Math100.0100.00.0
Veracode GenAI Code Security (Fall 2026)Cybersecurity73.073.00.0

Claude Opus 5.5vsGPT-6 Solnormalised score marginsData Nov 15, 2026Data: Mercor, Vals AI, Artificial Analysis and 65 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…