Atlas

Compare Models

GPT-6.1 Sol vs GPT-6 Astra

GPT-6 Astra leads by 1.0 capability points (≈95% interval −1.7 to +3.6): statistically tied · P(higher) ≈ 76%

Wins 236 of 352 shared benchmarks (77 lost, 39 tied)

Capability vs. relative cost
GPT-6.1 SolGPT-6 AstraClosed 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.

GPT-6.1 SolvsGPT-6 Astracapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins77 – 236 · 352 shared, 39 tied

Comparing matching setups only (7 benchmarks excluded) · suites counted once
← GPT-6.1 SolGPT-6 Astra →

GPT-6.1 SolvsGPT-6 Astrawins by categoryData Nov 15, 2026Data: Vals AI, Artificial Analysis, Mercor and 97 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is GPT-6 Astra minus GPT-6.1 Sol.

Clipping Bench v1.0Games57079514+40.9
EBR-bench v4 (card ban, single agent)Games54.376.2+21.9
Infrathon (mean rank score)Code54.475.2+20.8
Harvey's Legal Agent Benchmark - Energy Natural ResourcesLegal20.00.0−20.0
Harvey's Legal Agent Benchmark - Structured Finance SecuritizationLegal20.00.0−20.0
Harvey's Legal Agent Benchmark - Trusts Estates Private ClientLegal0.020.0+20.0
Clocktower BenchGames17641892+18.4
AA-PersonalityBench Imagination indexChat & Writing61.977.2+15.4
MatFlowBench Pass@1Science45.060.0+15.0
Short Form Bench v1.0Games29492803−14.5
Terminal-Bench-Science 0.1Science46.259.5+13.3
RWE-benchAgents25.037.0+12.0
EmbodiedLab (T0 success)Agents28.340.0+11.7
AA-PersonalityBench Organization indexChat & Writing81.069.5−11.5
ChessBench EloGames15621755+11.5
HardGeoBench Avg DistanceMultimodal1239908+10.2
EnigmaEvalGeneral QA59.569.7+10.2
SpireClimbGames23.834.0+10.2
ExploitBench - Internal Port (June-August 2026)Cybersecurity21.531.5+10.0
Terminal-Bench Science 0.1 (Vals)Science52.962.9+10.0
AA-PersonalityBench Warmth indexChat & Writing65.775.3+9.6
AA-PersonalityBench Conviction indexChat & Writing54.344.8−9.6
SnorkelManufacturingProfessional Work19.610.0−9.6
LisanBench Difficulty-WeightedGames40043113−9.5
ChessBench CoherenceGames87.597.0+9.5
Real-SWE 1.1Code47.256.0+8.8
HoneyBench v0.2 (hack rate)Safety59.550.8+8.7
HieroglyphBenchMultimodal31.440.0+8.6
BenchBench-CloningScience22.230.6+8.4
FrontierCS 2Code37.545.8+8.3
HPCT Refusal (CAIS)Safety42.050.0+8.0
ArXivMath 08/2026 (Native Harnesses)Math96.588.6−7.9
EnterpriseBench: CoreCraftProfessional Work50.342.6−7.7
AA-PersonalityBench Expressiveness indexChat & Writing50.558.1+7.7
Olam Social Poker EloGames15431550+7.6
Machiavelli Deception (CAIS)Safety78.876.3+7.6
SPU (Science Prediction Universe)Science0.30.4+7.5
HardGeoBench Country AccuracyMultimodal52.560.0+7.5
HardGeoBench <=100miMultimodal43.851.3+7.5
ExploitGym - release-reported percentage, cohort unspecifiedCybersecurity35.142.4+7.3
Humanity's Sixth SenseMultimodal46.653.6+7.0
MazeBench ASCIIGames9.016.0+7.0
AA-3D v1.0 EloCode11211168+6.9
MysteryMechanismScience46.453.2+6.8
TxBench Antibody Discovery V1 (100 evaluations)Science49.756.0+6.3
BuseyBench SVGMultimodal8.07.3−6.3
SRE Bench — Fully SolvedCode50.856.9+6.1
Chess Puzzles (task version 2.0)Games68.074.0+6.0
TaxCalcBench TY25Finance & Accounting38.044.0+6.0
RuneBenchGames6.97.3+5.9
FrontierML ($200 wallet)Agents38.744.5+5.9
Arena Image-to-WebDevCode16961722+5.7
RareDiseaseBench V1 (35 evaluations)Science27.633.3+5.7
Terminal-Bench 4.0 (Artificial Analysis)Code54.059.6+5.6
Design Creativity BenchCode63.765.5+5.5
NoCoT-Bench NCRI 15.2General QA149159+5.5
Vals Science IndexIndexes59.665.1+5.5
VoxelBench Text RatingGames25502625+5.3
AutomationBench 1.0.6 (provider-reported)Agents36.141.4+5.3
AutoRLBench v2 (solution level)Agents57.952.6−5.3
Agent Red Teaming k=100 (CAIS)Safety18.813.5+5.3
Finance Agent v2 - General Qualitative AnalysisFinance & Accounting63.268.4+5.2
DAYJOB: LawLegal31.836.9+5.1
IntPhys 2Multimodal72.377.4+5.1
Humanity's Last Exam DiamondGeneral QA58.863.8+5.0
TxBench-Antibody-Development V1 (60 evaluations)Science27.832.8+5.0
SciCode (main problems)Science14.619.6+5.0
Robo-benchAgents28.323.3−5.0
sudo L7Code28.933.9+5.0
SnorkelFinance 2.0Finance & Accounting6.611.6+5.0
ARC-AGI-3 (Semi-Private)General QA95.099.9+5.0
Web Clone BenchCode12.617.5+4.9
ALE-BenchCode29903128+4.9
Finance Agent v2 - ComparablesFinance & Accounting42.637.8−4.9
Snorkel Agentic Coding 2.0Code42.847.6+4.8
DAYJOB: AccountingFinance & Accounting25.230.0+4.8
FrontierCode 1.1 Extended - Pass RateCode65.970.6+4.8
AA-Robotics v0.1Agents64.669.4+4.8
ActiveVisionMultimodal71.876.5+4.7
FrontierCode 1.1 Extended - ScoreCode59.864.5+4.7
Code2Video-AA EloCode13691427+4.7
ChessBench AccuracyGames80.785.0+4.6
FrontierCode 2.0 - Pass RateCode43.047.6+4.5
Finance Agent v2 - PrecedentsFinance & Accounting25.830.3+4.5
Terminal-Bench 2.1Code85.489.9+4.5
FrontierCUA pass@1Agents46.742.2−4.4
ViBench 2.0Code64.760.3−4.4
BrokenArXiv 08/2026 (Native Harnesses)Math86.381.9−4.4
FrontierCode 2.0 - ScoreCode39.644.0+4.4
DAXBench v2Code92.696.8+4.3
Finance Agent v2 - General Quantitative AnalysisFinance & Accounting74.578.8+4.3
SheetBench-50Finance & Accounting60.564.7+4.2
LisanBench Path LengthGames87367393−4.2
ERQAMultimodal78.282.3+4.1
AA-LCR v1.1Search84.080.0−4.0
BullshitBench v2 LeaderboardSafety65.069.0+4.0
Concept Synth INDUCTION Challenge100 Round 1General QA90.094.0+4.0
DAYJOB: HealthcareHealth & Medicine7.611.6+4.0
MetagenomicsBench V1Science45.749.7+4.0
Mystery Game PuzzlesGames80.084.0+4.0
WeirdML v3Science38.342.2+4.0
LMArena Agent Arena Net Improvement (Steering Burden)Agents12.514.5+3.9
MedXpertQA MM Extended (Mercor held-out)Health & Medicine64.760.8−3.9
SnorkelUnderwrite 2.0Finance & Accounting25.021.2−3.8
τ³-Banking (Mercor run, bm25_grep retrieval)Agents43.046.7+3.7
Finance Agent v2 - AdjustmentsFinance & Accounting44.748.3+3.6
FrontierPhysics v0.1Science19.623.2+3.6
MindCube TinyMultimodal89.192.7+3.6
DAYJOB: FinanceFinance & Accounting18.021.5+3.5
SWE Atlas v1.1 - OverallCode65.969.4+3.5
DeepFinanceFinance & Accounting54.357.6+3.3
Finance Agent v2 - Disclosure AnalysisFinance & Accounting58.461.7+3.3
Code Migration v1.1Code77.380.6+3.3
Agent Initiative BenchmarkAgents8.79.0+3.2
BioSecBench-Refusal V1 Clean RerunsSafety37.640.8+3.2
AutoRLBench v2 (issue level)Agents43.846.9+3.1
Concept Synth INDUCTION Challenge64 Round 1General QA95.398.4+3.1
IOI (Vals v2)Code96.9100.0+3.1
PostTrainBench v1.2Indexes38.841.9+3.1
OpenWorldEngineer (Mean@8)Code41.244.2+3.0
GeneBench-Pro Public Release SubsetScience33.036.0+3.0
ScarfBench Smoke ScoreCode68.465.4−3.0
HiL-Bench (Human-in-Loop Benchmark)Agents62.365.3+3.0
SpatialViz-BenchMultimodal90.393.1+2.8
LMArena Code Arena WebDevCode17431764+2.8
HANDBOOK.md Agents v1.1Professional Work28.631.4+2.8
Mercor enterprise risk (benign-control completion rate)Safety100.097.2−2.8
Mercor enterprise risk (implicit-risk unsafe rate)Safety83.380.6+2.8
WARC-Bench (192-task test split)Agents86.389.1+2.8
TextQuests Game ProgressGames71.474.2+2.8
TissueMicroscopyBench V1 (49 evaluations)Science22.419.7−2.7
OncologyBench Strict (25 tasks, October 9 rubric)Health & Medicine24.021.3−2.7
Code MigrationCode65.167.7+2.6
BigFinanceBench private holdout (final-answer accuracy)Finance & Accounting52.054.6+2.6
BioSecBench-Function V1 Clean RerunsScience46.048.6+2.6
DAYJOB: Data ScienceProfessional Work18.721.3+2.6
Snorkel SWE-bench CLICode9.812.4+2.6
Extended NYT Connections 940 (Quadratic v1)Games95.598.1+2.6
MKA (Artificial Analysis)Professional Work47.750.2+2.5
Dual-Arm Sim (task completion)Agents0.02.5+2.5
HardGeoBench <=10miMultimodal35.037.5+2.5
ProgramBenchCode3.05.5+2.5
SUPERChemScience77.880.3+2.5
Artificial Analysis Engineering IndexIndexes59.461.8+2.4
FrontierMath Tier 4 (v2) - PrivateMath100.097.6−2.4
RiemannBench (November 2026 rescoring)Math79.281.6+2.4
Decision-Bench v1.1Professional Work49.151.5+2.4
GameBench 2Games92.094.3+2.4
Vision Arena OverallMultimodal12931283−2.3
Ramp Accounting BenchFinance & Accounting46.448.7+2.3
Humanity's Last ExamGeneral QA51.353.6+2.3
SWE-TogetherCode58.360.6+2.3
BioSecBench-Surveillance V1 Clean RerunsScience50.848.5−2.3
Artificial Analysis Intelligence Index v5.0.1Indexes54.656.9+2.2
Visa Agentic Commerce BenchmarkAgents62.264.4+2.2
CheatBenchSafety42.845.0−2.2
Frontier3DBenchCode66.063.9−2.2
LMArena Agent Arena Praise vs ComplaintAgents27.531.7+2.1
Finance Agent v2 - Earnings AnalysisFinance & Accounting71.569.5−2.0
WellnessBenchSafety66.064.0−2.0
AA-PersonalityBench Transparency indexChat & Writing92.494.3+1.9
AA-PersonalityBench Informativeness indexChat & Writing27.729.6+1.9
AA-PersonalityBench Slop indexChat & Writing14.316.2+1.9
OpenAI internal computer-use safety - adverse outcome rateSafety4.32.4+1.9
AutomationBench-AA 1.0.6Agents66.668.5+1.9
Artificial Analysis Legal IndexIndexes53.755.6+1.9
APEX-Agents 1.1Professional Work73.275.1+1.9
ArXivLean 09/2026Math35.937.7+1.9
DeepsecBench — Revalidated V1 (233 issues)Code39.537.6−1.9
Snorkel Digital Work IndexIndexes21.223.0+1.8
CyberBench v1.1 OverallCybersecurity39.341.1+1.8
APEX-SWECode46.948.6+1.7
Roboflow Vision Evals (October 2026, five tasks)Multimodal81.783.4+1.7
SimpleQA Verified (Epoch anti-abstention)General QA73.975.6+1.7
Artificial Analysis Healthcare & Medical IndexIndexes52.053.7+1.7
DataCenterBench Tier 1Agents38.340.0+1.7
Conceptual Reasoning IndexIndexes72.773.9+1.7
Excel Modeling Benchmark v1.1Finance & Accounting71.172.7+1.6
LMCAGeneral QA62.864.4+1.6
o11y-bench 1.1 Pass^3Agents87.585.9−1.6
Finance Agent v2 - Financial ModelingFinance & Accounting23.121.6−1.5
VCT-Refusal (CAIS)Safety16.117.6+1.5
Artificial Analysis Coding Agent Index v1.5Indexes60.161.6+1.5
FrontierMath ErdősMath4.42.9−1.5
MedScribeHealth & Medicine86.587.9+1.5
LMArena Agent Arena Bash ReliabilityAgents5.07.8+1.4
LMArena Text Arena EloChat & Writing14821474−1.4
Finance Agent v2.1Finance & Accounting61.860.4−1.4
HealthBench ProfessionalHealth & Medicine68.369.7+1.3
ChessBenchGames68.770.0+1.3
Professional Reasoning Benchmark - FinanceFinance & Accounting46.247.5+1.3
MMMU ProMultimodal85.186.4+1.3
WorkflowArenaAgents29.530.8+1.3
MMMU-Pro Extended (Mercor held-out)Multimodal61.360.0−1.3
Android Bench 2.0Code29.328.0−1.3
AA-AnalystAgent (pass^5)Professional Work50.051.3+1.2
BrowseComp (Mercor 130-question subset)Search95.494.2−1.2
GameDevBenchGames70.068.8−1.2
Hebbia Financial Services BenchmarkFinance & Accounting73.572.3−1.2
SnorkelRevOpsProfessional Work12.110.9−1.2
ForecastBench Preliminary Dataset Brier IndexForecasting61.762.9+1.2
OSWorld 2 (Aug 8) offline - reported partial reward, step limit unspecifiedAgents71.472.6+1.2
GDPxProfessional Work8.49.6+1.2
LMArena Agent Arena Confirmed SuccessAgents11.714.0+1.2
Artificial Analysis Finance & Accounting IndexIndexes56.257.4+1.1
DeepSWE v1.1Code75.274.1−1.1
Medical Long Context Reasoning (MLCR-AA)Health & Medicine33.935.0+1.1
AA-Omniscience IndexIndexes41.543.7+1.1
MedXpertQA MM (200-question set)Health & Medicine85.786.8+1.1
Roboflow Vision Evals (September 2026 protocol)Multimodal85.586.6+1.1
Finance Agent v2 - Market AnalysisFinance & Accounting64.565.5+1.1
NoCoT-Bench NCKI (kspine v3)General QA130131+1.1
Tax Agent BenchFinance & Accounting62.363.3+1.0
Financial Coaching BenchFinance & Accounting68.567.5−1.0
BrokenArXiv 05/2026Math90.091.0+1.0
CrossSearch (pass^3)Search59.060.0+1.0
Toloka Arena (v3)Agents66.067.0+1.0
Legal Research BenchLegal38.539.4+1.0
Artificial Analysis Coding Agent Index v2.1Indexes52.253.1+0.9
SM BenchSafety69.069.9+0.9
BrokenArXiv 06/2026Math100.099.1−0.9
Excel Modeling BenchmarkFinance & Accounting70.871.7+0.9
BioSecBench-RespiratoryVirus-Surveillance V1 (110 evaluations)Science12.413.3+0.9
GDPval-AA v2.1Professional Work16071587−0.9
GDP.xlsxProfessional Work22.022.9+0.9
Artificial Analysis Economics IndexIndexes56.357.2+0.9
CritPt v2 (agentic, avg@5)Science82.683.4+0.9
Hermes IndexIndexes57.156.3−0.8
Artificial Analysis Intelligence Index v4.3.2Indexes51.852.7+0.8
SpeciEval (Pooled English Runs)Safety97.496.5−0.8
ARC-AGI-2General QA94.295.0+0.8
HLAB-Vals-1.1Legal6.37.1+0.8
SpeciEval (Refusal-Aware Retries, Historical)Safety97.496.5−0.8
ACCoRDGeneral QA80.981.7+0.8
MentalHealthBenchSafety57.958.7+0.8
HealthBench Hard (updated implementation)Health & Medicine33.434.2+0.8
DAYJOB: TaxFinance & Accounting4.04.8+0.8
DTBenchGeneral QA96.597.3+0.8
MissionReportProfessional Work61.061.8+0.8
VIALSMultimodal49.149.9+0.8
SWE-Bench Pro V2 — FullCode96.196.9+0.8
AA-Briefcase v1.1 EloAgents15551569+0.8
Artificial Analysis Intelligence Index v5.0Indexes54.555.3+0.7
NeuroscienceBench V1 (91 evaluations)Science22.022.7+0.7
LiveBench 2.0Agents85.486.1+0.7
LMArena Agent Arena Steering BurdenAgents13.414.8+0.7
Agents' Last Exam Overall ScoreAgents58.659.3+0.7
ComplexConstraints (November 2026 rescoring)Chat & Writing70.170.8+0.7
SimpleBenchGeneral QA82.983.6+0.7
SciCode 1.0.1 (Artificial Analysis)Code55.856.5+0.7
Legacy-Bench (Fireworks, avg@3)Code55.055.7+0.7
Harvey LAB: In-House Contracting (held-out)Legal4.75.3+0.7
Bug Hunt BenchCode42.242.9+0.6
BioSimBench 1.0Science15.315.9+0.6
SWE Atlas - Codebase QnACode48.447.8−0.6
Snorkel WorkplaceAgentsProfessional Work15.014.4−0.6
ClockBench AccuracyMultimodal74.475.0+0.6
SpatialBench-Long V1 Clean RerunsScience39.939.3−0.6
TxBench Oligo Discovery V1 (113 evaluations)Science52.551.9−0.6
ScBench-Long V1 Clean RerunsScience26.727.3+0.6
VariantBench V1 Clean RerunsScience46.046.6+0.6
Artificial Analysis Coding Agent Index v2.0Indexes54.154.7+0.5
Westworld Finance Diligence BenchFinance & Accounting55.054.5−0.5
Bedside BenchHealth & Medicine93.993.4−0.5
MacroscopeBenchCode78.578.0−0.5
MMLU ProGeneral QA91.091.5+0.5
Wiz Cyber Model Arena v3.9Cybersecurity70.770.2−0.5
LMArena Agent Arena Bash RecoveryAgents4.55.5+0.5
Political Science LLM BenchmarkProfessional Work68.367.9−0.5
Professional Reasoning Benchmark - LegalLegal47.948.4+0.5
Vals RSI Index v1.1 (four tasks)Agents26.627.1+0.5
TextQuests Average Cumulative HarmSafety29.329.2+0.4
AdvancedIFChat & Writing82.482.0−0.4
APEX-Accounting 1.1Finance & Accounting64.364.7+0.4
Harvey LAB (Mercor run, undisclosed ~213-task cohort)Legal8.89.2+0.4
HealthBench Hard (length-adjusted)Health & Medicine36.236.6+0.4
FACTS Benchmark SuiteGeneral QA75.375.7+0.4
Artificial Analysis Progress Index v1Indexes348350+0.4
VulcanBench Frontier v4Code88.889.2+0.4
BioMysteryBench (Vals Terminus 2)Science79.679.3−0.4
Search ArenaSearch12301231+0.4
Artificial Analysis Strategy & Ops IndexIndexes57.658.0+0.4
MedCodeHealth & Medicine48.848.5−0.4
Finance Agent v2 - All PassFinance & Accounting39.940.3+0.3
FrontierBioMLScience29.128.8−0.3
TxBench Preclinical Pharmacology V1 Clean RerunsScience59.759.3−0.3
ScBench V1 Clean RerunsScience63.663.9+0.3
Vals Legal IndexIndexes68.768.4−0.3
Harvey HLAB — Criteria Pass RateLegal87.887.5−0.3
SpatialBench V1 Clean RerunsScience67.367.5+0.3
Hermes BenchAgents73.973.7−0.3
Epiq LAER CorporateBenchProfessional Work82.282.0−0.2
Epoch Capabilities Index (ECI)Indexes169169+0.2
GBENCH One-shot Coding GScoreGames79.079.2+0.2
SWE-bench Multimodal (Vals v1, 480 Tasks)Code43.543.3−0.2
CVDP (517-problem set)Code55.154.9−0.2
GDP.pdf (Artificial Analysis, All-pass)Professional Work32.032.2+0.2
MASKSafety98.398.5+0.2
SnorkelLegalLegal20.720.9+0.2
τ³-Banking (Fireworks, banking_knowledge)Agents44.744.8+0.2
SAGEMath46.546.4−0.2
Vals Finance IndexIndexes55.855.6−0.2
Arena Alignment IndexSafety87.987.8−0.2
Prophet Arena Arena Score (Agentic, OpenCode)Forecasting86.086.2+0.2
DistressBenchSafety74.274.3+0.1
Immigration BenchLegal67.867.7−0.1
Prophet Arena Arena Score (Fixed Context)Forecasting84.884.7−0.1
Ruby LLM Benchmarks — Program FixerCode75.075.1+0.1
PCT Political Inconsistency (CAIS)Safety37.337.4−0.1
CharXiv (descriptive and reasoning combined)Multimodal96.096.1+0.1
Gray Swan IPI Attack Success (k=15)Safety8.68.5+0.1
LLM Chess EloGames16121614+0.1
HealthBenchHealth & Medicine56.956.90.0
LMArena Agent Arena Tool HallucinationAgents1.01.00.0
AIME 2025Math100.0100.00.0
ARC-AGI-1 Semi-PrivateGeneral QA98.598.50.0
ArXivMath 05/2026Math95.095.00.0
ArXivMath 06/2026Math94.494.40.0
BioSecBench-Structure V1 (58 evaluations)Science21.321.30.0
CharXiv Extended (Mercor held-out)Multimodal85.185.10.0
CritPt (avg@5)Science31.731.70.0
DualEntry Accounting AI BenchmarkFinance & Accounting81.281.20.0
EpiBench V1 Clean RerunsScience28.328.30.0
FrontierMath Tiers 1-3 (v2) - PrivateMath93.793.70.0
Furniture Assembly BenchmarkMultimodal80.080.00.0
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&ALegal20.020.00.0
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 - Healthcare Life SciencesLegal0.00.00.0
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 - TaxLegal0.00.00.0
Harvey's Legal Agent Benchmark - White Collar Defense InvestigationsLegal0.00.00.0
Mercor enterprise risk (explicit-risk unsafe rate)Safety0.00.00.0
ObviousBenchGeneral QA100.0100.00.0
OTIS Mock AIME 2024-2025Math100.0100.00.0
ProofBench v1.1Math99.099.00.0
SWE-Marathon v1.1Code42.542.50.0
Terminal-Bench 4.0Code58.258.20.0

GPT-6.1 SolvsGPT-6 Astranormalised score marginsData Nov 15, 2026Data: Vals AI, Artificial Analysis, Mercor and 97 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…