Atlas

Compare Models

Claude Opus 5 vs Claude Opus 5.5

Claude Opus 5.5 leads by 3.7 capability points (≈95% interval +1.1 to +6.2): a clear lead over Claude Opus 5 · P(higher) ≈ 100%

Wins 200 of 281 shared benchmarks (55 lost, 26 tied)

Capability vs. relative cost
Claude Opus 5Claude Opus 5.5Closed 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 5vsClaude Opus 5.5capability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins55 – 200 · 281 shared, 26 tied

Comparing matching setups only (18 benchmarks excluded) · suites counted once
← Claude Opus 5Claude Opus 5.5 →

Claude Opus 5vsClaude Opus 5.5wins by categoryData Nov 15, 2026Data: Vals AI, Mercor, Artificial Analysis and 87 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Claude Opus 5.5 minus Claude Opus 5.

Time Horizon Index: KSPAgents18.891.3+72.5
VCT-Refusal (CAIS)Safety29.490.4+61.0
Machiavelli Deception (CAIS)Safety96.577.5+57.4
Anthropic ECI (AECI), November 2026 fitIndexes169175+47.2
TasteVal (compute multiplier)Science1.32.3+44.7
TxBench Oligo Discovery V1 Clean RerunsScience45.61.1−44.4
Concept Synth INDUCTION Challenge100 Round 1General QA24.065.0+41.0
Concept Synth INDUCTION Challenge64 Round 1General QA32.871.9+39.1
CheatBenchSafety45.37.2+38.1
TxBench Antibody Discovery V1 Clean RerunsScience41.96.8−35.0
YC-BenchGames19115922873725+33.6
Terminal-Bench-Science 0.1Science28.661.9+33.3
BioSecBench-Refusal V1 Clean RerunsSafety31.80.0−31.8
BioSecBench-Function V1 Clean RerunsScience42.573.6+31.2
ChessBench EloGames8261349+31.1
Into the Unknown Integrity Node Pass RateScience59.590.4+30.9
KernelBench CUDA Normalized (RTX PRO 6000)Code45.576.3+30.8
ActiveVisionMultimodal10.641.2+30.6
TxBench Preclinical Pharmacology V1 Clean RerunsScience59.730.7−29.0
LisanBench Path LengthGames1542824029+26.7
EBR-bench v4 (card ban, single agent)Games45.771.4+25.7
Terminal-Bench Science 0.1 (Vals)Science22.948.6+25.7
SPU (Science Prediction Universe)Science0.2-0.3−24.8
Chess Puzzles (task version 2.0)Games43.067.0+24.0
LMArena Agent Arena Net Improvement (Steering Burden)Agents7.219.5+23.2
CrossSearch (pass^3)Search34.057.0+23.0
HieroglyphBenchMultimodal14.637.4+22.8
Furniture Assembly BenchmarkMultimodal60.883.3+22.5
VoxelBench Text RatingGames23002613+22.3
TaxCalcBench TY25Finance & Accounting20.042.0+22.0
FrontierMath Tier 4 (v2) - PrivateMath73.295.0+21.8
FrontierCode 2.0 - Pass RateCode30.451.8+21.4
Blueprint-Bench 2Multimodal0.30.5+20.8
Search ArenaSearch12391213−20.6
FrontierCode 2.0 - ScoreCode27.247.3+20.1
CadQueryEval (Corrected Geometry)Code80.0100.0+20.0
Harvey's Legal Agent Benchmark - Corporate M&ALegal20.00.0−20.0
Harvey's Legal Agent Benchmark - Structured Finance SecuritizationLegal20.00.0−20.0
LAB-Bench LitQA2 (SecureBio)Science51.471.1+19.7
Agent Red Teaming k=100 (CAIS)Safety43.324.4+18.9
FrontierPhysics v0.1Science9.527.4+17.9
ZeroBench Main Mean pass@1 (five samples)Multimodal17.235.0+17.8
ZeroBench (Main questions, pass@5)Multimodal26.043.0+17.0
HardGeoBench Avg DistanceMultimodal1218669+17.0
MMMU-Pro Extended (Mercor held-out)Multimodal34.351.0+16.7
BrowseComp Extended (Mercor held-out)Search44.060.5+16.5
DAYJOB: HealthcareHealth & Medicine8.424.7+16.3
ChessBench CoherenceGames63.880.1+16.3
LMArena Code Arena WebDevCode16911814+16.1
SWE-bench Verified Extended (Mercor held-out)Code82.066.0−16.0
VIALSMultimodal24.440.2+15.8
ProgramBench (Mercor 100-task subset)Code4.320.0+15.7
Hedge-BenchFinance & Accounting29.044.2+15.2
HardGeoBench <=100miMultimodal46.361.3+15.0
RuneBenchGames5.76.7+14.7
WCB (SecureBio)Science68.854.5−14.3
Humanity's Sixth SenseMultimodal30.544.6+14.1
LMArena Agent Arena Praise vs ComplaintAgents12.240.3+14.1
ZeroBench Main Questions pass^5Multimodal11.025.0+14.0
Bug Hunt BenchCode25.739.7+14.0
EnigmaEvalGeneral QA43.957.8+13.9
SpatialViz-BenchMultimodal75.889.7+13.9
EpiBench V1 Clean RerunsScience25.239.0+13.8
Clipping Bench v1.0Games7342018+13.8
MedCodeHealth & Medicine63.649.8−13.8
HardGeoBench Country AccuracyMultimodal53.867.5+13.8
Into the Unknown Direction Node Pass RateScience66.880.0+13.2
Terminal-Bench 4.0 (Artificial Analysis)Code46.559.6+13.1
Olam Social Poker EloGames15181506−13.0
Toloka Arena (v3)Agents62.075.0+13.0
MaxxBenchCode63.576.4+12.9
LAB-Bench ProtocolQA (SecureBio)Science81.194.0+12.9
GDPval Extended (Mercor held-out rubric score)Professional Work62.975.8+12.9
FrontierSWE V2 Worst@5Code39.152.0+12.9
ComplexConstraintsChat & Writing37.349.9+12.6
DAYJOB: FinanceFinance & Accounting11.323.9+12.6
ExploitGymCybersecurity22.034.5+12.5
ReactBench (39 tasks)Code42.154.4+12.3
WeirdML v3Science19.031.2+12.2
Eval Connections One-Shot with TrapsGames74.086.0+12.0
Mystery Game PuzzlesGames59.071.0+12.0
Harvey HLAB — Criteria Pass RateLegal87.775.8−11.9
LLM Chess EloGames12851575+11.9
ClockBench AccuracyMultimodal65.577.2+11.7
SpatialBench-Long V1 Clean RerunsScience30.919.6−11.3
CursorBench 4.0Code46.657.8+11.2
LABBench2Science84.273.1−11.1
Roboflow Vision Evals (October 2026, five tasks)Multimodal70.982.0+11.1
τ³-Banking (Mercor run, bm25_grep retrieval)Agents44.055.0+11.0
Terminal-Bench 4.0Code53.964.8+10.9
BioSecBench-RespiratoryVirus-Surveillance V1 (110 evaluations)Science14.625.4+10.9
FrontierSWE V2 Best@5Code60.771.5+10.9
ScBench-Long V1 Clean RerunsScience24.413.6−10.8
SciCode 1.0.1 (Artificial Analysis)Code56.466.9+10.5
Design Arena Overall Code CategoriesCode13211384+10.4
GlobeBenchGeneral QA15.926.3+10.4
Drone-BenchAgents86.997.3+10.4
BioSecBench-Structure V1 (58 evaluations)Science20.731.0+10.3
CritPt v2 (agentic, avg@5)Science71.481.7+10.3
BullshitBench v2 LeaderboardSafety73.063.0−10.0
HardGeoBench <=10miMultimodal32.542.5+10.0
SciCode (main problems)Science17.327.3+10.0
VCT (SecureBio)Science51.461.4+10.0
GameBench 2Games84.894.8+10.0
Humanity's Last Exam DiamondGeneral QA44.854.6+9.8
Humanity's Last Exam RMS Calibration Error (CAIS)General QA23.613.8+9.8
MetagenomicsBench V1Science53.343.7−9.7
MetagenomicsBench V0Science53.343.7−9.6
LisanBench Difficulty-WeightedGames50685963+9.6
DAXBench v2Code86.295.7+9.6
Terminal-Bench 2.1 Extended (Mercor held-out)Code38.047.5+9.5
dfbench Detection RecallCybersecurity45.454.9+9.5
BrokenArXiv 06/2026Math90.7100.0+9.3
ChessBench AccuracyGames71.079.3+8.9
MindCube TinyMultimodal84.393.1+8.8
HackerRank-BenchCode42.651.2+8.7
MBCT (SecureBio)Science60.268.7+8.5
LVBenchMultimodal75.483.7+8.3
GPQA-Bio (SecureBio)Science82.790.9+8.2
AA-Briefcase v1.1 EloAgents16541801+8.0
GeneBench-Pro Public Release SubsetScience22.030.0+8.0
PLE (Physical-Engineering's Last Exam)Professional Work41.949.9+8.0
SimpleBenchGeneral QA80.688.4+7.8
MineBench — Bradley-Terry RatingGames20082134+7.6
PCT Political Inconsistency (CAIS)Safety36.729.1+7.6
Arena Alignment IndexSafety75.683.2+7.6
APEX-AccountingFinance & Accounting54.561.8+7.3
Agents' Last Exam (ALE)Agents30.938.2+7.2
Medical Long Context Reasoning (MLCR-AA)Health & Medicine59.466.7+7.2
GDP.pdf (Artificial Analysis, All-pass)Professional Work21.628.8+7.2
Roboflow Vision Evals (September 2026 protocol)Multimodal78.385.5+7.2
RSI-Exam 0.1 (Full 88)Agents0.50.5+7.2
MedXpertQA MM Extended (Mercor held-out)Health & Medicine64.171.2+7.1
LMArena Agent Arena Steering BurdenAgents4.418.6+7.1
τ³-Banking (Fireworks, banking_knowledge)Agents49.156.0+6.9
EQ-Bench 4Chat & Writing13651340−6.7
BenchCAD - geometric overlapMultimodal89.996.2+6.3
Artificial Analysis Coding Agent Index v1.5Indexes59.766.0+6.3
GDPval-AA v2.1Professional Work17291866+6.1
Physical-Engineering's Last Exam (hidden test)Professional Work42.648.6+6.0
MedXpertQA MM (200-question set)Health & Medicine79.885.7+5.9
Extended NYT Connections 940 (Quadratic v1)Games94.388.5−5.8
DeepsecBench — Revalidated V1Code32.426.7−5.8
GBENCH One-shot Coding GScoreGames68.173.9+5.8
DeepsecBench — Revalidated V1 (233 issues)Code32.326.6−5.7
EnterpriseBench: CoreCraftProfessional Work68.774.4+5.7
FrontierMath Tiers 1-3 (v2) - PrivateMath85.691.2+5.6
Terminal-Bench 2.1Code89.183.5−5.6
BuseyBench SVGMultimodal6.36.8+5.5
Bio Capabilities Index (SecureBio)Indexes178171−5.5
DistressBenchSafety70.275.6+5.4
TextQuests Average Cumulative HarmSafety22.624.2−5.3
LMArena Agent Arena Confirmed SuccessAgents10.220.7+5.3
Vals RSI Index v1.1 (historical five tasks)Agents32.137.1+5.0
CUA-benchAgents9.014.0+5.0
Chartography (with tools)Multimodal84.089.0+5.0
SnorkelLegalLegal26.431.4+5.0
SWE-TogetherCode63.868.8+5.0
CharXiv Extended (Mercor held-out)Multimodal79.884.7+4.9
SUPERChemScience74.779.6+4.9
AA-Omniscience IndexIndexes37.146.4+4.7
Senior SWE-BenchCode34.739.3+4.6
ACCoRDGeneral QA83.888.4+4.6
ALE-BenchCode26222748+4.4
LiveBench 2.0Agents82.186.5+4.4
Conceptual Reasoning IndexIndexes74.777.9+4.4
SnorkelUnderwrite 2.0Finance & Accounting18.322.7+4.4
SpatialBench V1 Clean RerunsScience68.472.8+4.3
BioSecBench-Surveillance V1 Clean RerunsScience51.555.8+4.3
APEX-Agents 1.1Professional Work77.081.3+4.3
Vals RSI Index v1.1 (four tasks)Agents33.037.3+4.3
ERQAMultimodal70.875.1+4.3
Creative Writing v3 EloChat & Writing21352051−4.3
SM BenchSafety85.881.8−4.0
TextQuests Game ProgressGames53.357.3+4.0
MASKSafety93.489.4−4.0
BrowseComp (Mercor 130-question subset)Search84.688.5+3.9
APEX-SWECode63.767.6+3.9
LMCAGeneral QA64.568.2+3.8
PostTrainBench Average (v1.1)Indexes45.649.3+3.7
FORTRESS Average Risk Score (ARS)Safety16.813.2+3.7
SAGEMath49.445.8−3.6
Epoch Capabilities Index (ECI)Indexes164169+3.4
SWE Atlas Codebase QnA Extended (Mercor held-out)Code40.043.3+3.3
LLM Debate BenchmarkChat & Writing17241701−3.3
MacroscopeBenchCode69.973.1+3.2
IntPhys 2Multimodal65.168.3+3.1
Humanity's Last Exam (Text-Only)General QA54.457.5+3.1
VariantBench V1 Clean RerunsScience52.849.7−3.1
AutoResearchExam (hidden-test AUARC)Agents0.60.6+3.0
FrontierMath ErdősMath0.02.9+2.9
ARC-AGI-2General QA90.493.3+2.9
Humanity's Last Exam (with tools)General QA64.867.7+2.9
LMArena Text Arena EloChat & Writing14901506+2.8
HealthBenchHealth & Medicine67.164.3−2.8
DeepResearch Bench II (Mercor 130-task set)Search56.153.3−2.8
APEX-1 (Mercor, mean score)Professional Work68.665.8−2.8
AURORA Multilingual Terminal-BenchCode73.776.5+2.8
GDPval (Mercor 213-task rubric score)Professional Work85.688.3+2.7
AA-LCR v1.1Search82.084.7+2.7
LMArena Agent Arena Bash ReliabilityAgents9.815.1+2.7
MMMU ProMultimodal84.086.6+2.6
Snorkel SWE-bench CLICode14.016.6+2.6
CritPt (avg@5)Science29.131.7+2.6
RCM-Bench-1Health & Medicine92.790.1−2.5
SAB RankingForecasting-0.1-0.1+2.5
AA-AnalystAgent (pass^5)Professional Work53.856.3+2.5
Excel Modeling BenchmarkFinance & Accounting73.675.9+2.4
AdvancedIFChat & Writing80.582.8+2.3
Ruby LLM Benchmarks — Program FixerCode79.877.5−2.3
Harvey LAB-AA v1.0 (criterion pass rate)Legal93.591.2−2.2
SnorkelFinance 2.0Finance & Accounting16.518.6+2.1
HiL-Bench (Human-in-Loop Benchmark)Agents57.059.0+2.0
RedlineBenchLegal44.042.0−2.0
HealthBench ProfessionalHealth & Medicine73.471.4−2.0
CVDP (517-problem set)Code62.264.1+1.9
Snorkel Digital Work IndexIndexes21.823.6+1.9
CharXiv (descriptive and reasoning combined)Multimodal94.696.4+1.8
Learn2Play BenchGames81.880.1−1.7
Longform Creative WritingChat & Writing86.384.6−1.7
MMLU-Bio (SecureBio)Science91.192.7+1.6
RiemannBenchMath68.069.6+1.6
Snorkel Agentic Coding 2.0Code38.940.4+1.5
Creative Writing v3 Rubric ScoreChat & Writing16.316.0−1.4
MMLU ProGeneral QA92.093.4+1.4
Epiq LAER CorporateBenchProfessional Work74.075.4+1.4
OncologyBench Strict (25 tasks, October 9 rubric)Health & Medicine36.037.3+1.3
BioTIER RefusalSafety100.098.7−1.3
SWE-Marathon v1.1Code46.945.6−1.2
ScBench V1 Clean RerunsScience57.656.4−1.2
FrontierCode 1.1 Extended - Pass RateCode69.670.8+1.2
OTIS Mock AIME 2024-2025Math98.9100.0+1.1
Harvey LAB held-out set (Fireworks run)Legal7.46.3−1.1
micro1 PharmaBenchHealth & Medicine9.710.8+1.1
DTBenchGeneral QA97.998.9+1.1
DeepSWE v1.1Code73.274.2+1.0
SAB ResearchFinance & Accounting96.297.2+1.0
ARC-AGI-1 Semi-PrivateGeneral QA97.598.5+1.0
BigFinanceBench private holdout (final-answer accuracy)Finance & Accounting49.248.2−1.0
Harvey LAB: In-House Contracting (held-out)Legal2.03.0+1.0
ProofBench v1.1Math99.0100.0+1.0
SnorkelManufacturingProfessional Work9.410.4+1.0
ATLAS-Finance v1.0Finance & Accounting12.311.4−0.9
LMArena Agent Arena Bash RecoveryAgents13.615.0+0.7
ScarfBench Smoke ScoreCode69.870.5+0.7
SpeciEval (Refusal-Aware Retries, Historical)Safety87.987.2−0.7
SpeciEval (Pooled English Runs)Safety87.987.2−0.7
Snorkel WorkplaceAgentsProfessional Work14.914.4−0.5
MedScribeHealth & Medicine91.091.4+0.4
SnorkelRevOpsProfessional Work12.212.5+0.3
AA-LCR Extended (Mercor held-out)Search73.974.2+0.3
ForecastBench Preliminary Dataset Brier IndexForecasting62.062.2+0.2
LLM Creative Story WritingChat & Writing3.73.7−0.2
LMArena Agent Arena Tool HallucinationAgents0.90.7−0.1
Biopharma Bench V0.1Health & Medicine67.667.5−0.1
AIME 2025Math100.0100.00.0
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 - 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 - Energy Natural ResourcesLegal20.020.00.0
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 - Healthcare Life SciencesLegal20.020.00.0
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 - 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
HPCT Refusal (CAIS)Safety100.0100.00.0
Into the UnknownScience0.00.00.0
ObviousBenchGeneral QA100.0100.00.0

Claude Opus 5vsClaude Opus 5.5normalised score marginsData Nov 15, 2026Data: Vals AI, Mercor, Artificial Analysis and 87 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…