Atlas

Compare Models

GPT-5.6 Terra vs GPT-6.1 Sol

GPT-6.1 Sol leads by 5.6 capability points (≈95% interval +3.3 to +7.9): a clear lead over GPT-5.6 Terra · P(higher) ≈ 100%

Wins 162 of 208 shared benchmarks (23 lost, 23 tied)

Capability vs. relative cost
GPT-5.6 TerraGPT-6.1 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.

GPT-5.6 TerravsGPT-6.1 Solcapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins23 – 162 · 208 shared, 23 tied

Comparing matching setups only (13 benchmarks excluded) · suites counted once
← GPT-5.6 TerraGPT-6.1 Sol →

GPT-5.6 TerravsGPT-6.1 Solwins by categoryData Nov 15, 2026Data: Vals AI, Mercor, Arena and 44 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is GPT-6.1 Sol minus GPT-5.6 Terra.

Concept Synth INDUCTION Challenge100 Round 1General QA25.090.0+65.0
Concept Synth INDUCTION Challenge64 Round 1General QA37.595.3+57.8
ARC-AGI-3 (Semi-Private)General QA0.852.7+51.9
Mystery Game PuzzlesGames35.080.0+45.0
Arena Code WebDev — FullstackCode15651735+44.6
Terminal-Bench-Science 0.1Science13.858.1+44.3
Terminal-Bench Science 0.1 (Vals)Science10.052.9+42.9
BrowseComp Extended (Mercor held-out)Search25.367.7+42.4
SRE Bench — Fully SolvedCode9.550.8+41.2
GameBench 2Games53.592.0+38.5
ChessBench EloGames9251562+37.9
Arena Code WebDev — HTMLCode15231793+37.7
Terminal-Bench 4.0Code21.558.2+36.7
EnigmaEvalGeneral QA23.159.5+36.4
Agent Red Teaming k=100 (CAIS)Safety54.618.8+35.8
Olam Social Poker EloGames15111543+34.8
Arena Code WebDev — Reference-Based DesignCode15161779+34.7
TreasuryQA BenchProfessional Work16.751.4+34.7
Arena Code WebDev — SimulationsCode15261829+34.3
Arena Image-to-WebDevCode15391696+34.2
Arena Code WebDev — Content Creation ToolsCode15031716+33.7
SWE-bench Verified Extended (Mercor held-out)Code78.745.3−33.4
NoCoT-Bench NCRI 15.2General QA89.0149+33.0
CyberBench v1.1 OverallCybersecurity72.139.3−32.8
Arena Code WebDev — GamingCode15481833+31.9
Arena Code WebDev — FrontendCode15211767+31.6
Arena Code WebDev — ReactCode15111747+31.5
TextQuests Game ProgressGames40.571.4+30.9
LisanBench Difficulty-WeightedGames11974004+30.1
Humanity's Last Exam DiamondGeneral QA28.858.8+30.0
ALE-BenchCode21542990+29.6
FrontierMath Tier 4 (v2) - PrivateMath70.7100.0+29.3
Gray Swan IPI Attack Success (k=15)Safety37.38.6+28.7
LMArena Agent Arena Net Improvement (Steering Burden)Agents-1.812.5+27.0
MindCube TinyMultimodal62.689.1+26.5
Furniture Assembly BenchmarkMultimodal54.280.0+25.8
Arena Code WebDev — Brand & MarketingCode15301707+25.6
SpatialViz-BenchMultimodal65.290.3+25.1
GeneBench-Pro Public Release SubsetScience8.033.0+25.0
ProofBench v1.1Math74.099.0+25.0
TextQuests Average Cumulative HarmSafety22.229.3−24.1
Humanity's Sixth SenseMultimodal22.746.6+23.9
Arena Code WebDev — Data & AnalyticsCode14961654+23.7
FACTS Benchmark SuiteGeneral QA52.075.3+23.3
HardGeoBench Avg DistanceMultimodal19911239+23.2
Arena Code WebDev — Consumer ProductCode15051685+22.8
Artificial Analysis Intelligence Index v5.0.1Indexes32.354.6+22.3
Chess Puzzles (task version 2.0)Games47.068.0+21.0
RuneBenchGames5.56.9+20.8
AA-Omniscience IndexIndexes0.141.5+20.7
Terminal-Bench 4.0 (Artificial Analysis)Code35.456.1+20.7
DeepsecBench — Revalidated V1 (233 issues)Code19.139.5+20.4
Harvey's Legal Agent Benchmark - Corporate M&ALegal0.020.0+20.0
Harvey's Legal Agent Benchmark - Energy Natural ResourcesLegal0.020.0+20.0
Design Arena Overall Code CategoriesCode12381356+19.6
BuseyBench SVGMultimodal6.18.0+18.7
NoCoT-Bench NCKI (kspine v3)General QA104130+17.9
ChessBench CoherenceGames69.787.5+17.8
Code MigrationCode47.865.1+17.3
Extended NYT Connections 940 (Quadratic v1)Games78.495.5+17.1
HPCT Refusal (CAIS)Safety25.042.0+17.0
LisanBench Path LengthGames33038736+16.9
FrontierCode 2.0 (Artificial Analysis)Code23.940.6+16.7
DAYJOB: FinanceFinance & Accounting1.818.0+16.2
FrontierPhysics v0.1Science3.619.6+16.0
APEX-SWECode31.146.9+15.8
Roboflow Vision Evals (October 2026, five tasks)Multimodal66.081.7+15.7
GBENCH One-shot Coding GScoreGames63.379.0+15.7
Humanity's Last ExamGeneral QA35.951.3+15.4
SWE Atlas Codebase QnA Extended (Mercor held-out)Code16.031.3+15.3
DeepSWE v1.1Code60.275.2+15.0
LMArena Agent Arena Praise vs ComplaintAgents-1.927.5+14.7
ERQAMultimodal63.678.2+14.6
Vibe Code Bench v1.1Code74.688.9+14.3
MMMU-Pro Extended (Mercor held-out)Multimodal47.461.3+13.9
MedXpertQA MM Extended (Mercor held-out)Health & Medicine51.064.7+13.7
LLM Chess EloGames12931612+13.1
SUPERChemScience64.977.8+12.9
CritPt v2 (agentic, avg@5)Science69.782.6+12.9
AA-Briefcase v1.1 EloAgents13331555+12.1
Bug Hunt BenchCode30.542.2+11.7
Roboflow Vision Evals (September 2026 protocol)Multimodal73.885.5+11.7
MentalHealthBenchSafety46.257.9+11.7
IntPhys 2Multimodal61.372.3+11.0
BullshitBench v2 LeaderboardSafety54.065.0+11.0
HieroglyphBenchMultimodal20.531.4+10.9
YC-BenchGames636144936267+10.5
LiveBench 2.0Agents75.085.4+10.4
ARC-AGI-2General QA83.994.2+10.3
Agents' Last Exam Overall ScoreAgents48.558.6+10.1
HardGeoBench <=10miMultimodal25.035.0+10.0
MedXpertQA MM (200-question set)Health & Medicine75.785.7+10.0
SWE-Marathon v1.1Code32.542.5+10.0
SM BenchSafety59.069.0+10.0
Artificial Analysis Intelligence Index v4.3.2Indexes42.151.8+9.8
Conceptual Reasoning IndexIndexes65.772.7+9.6
BrowseComp (Mercor 130-question subset)Search85.895.4+9.6
SWE-bench MultilingualCode94.484.9−9.5
SWE-bench VerifiedCode77.887.3+9.5
LMArena Agent Arena Confirmed SuccessAgents-7.211.7+9.5
ChessBench AccuracyGames71.980.7+9.4
Terminal-Bench 2.1 Extended (Mercor held-out)Code36.445.8+9.4
IOI (Vals v2)Code87.696.9+9.3
PCT Political Inconsistency (CAIS)Safety46.237.3+8.9
Artificial Analysis Progress Index v1Indexes324348+8.6
Finance Agent v2 - Disclosure AnalysisFinance & Accounting66.958.4−8.5
AML BenchFinance & Accounting62.370.6+8.3
Vals Index v2.1Indexes53.161.2+8.1
Harvey HLAB — Criteria Pass RateLegal79.887.8+8.0
LMCAGeneral QA55.062.8+7.8
FrontierMath Tiers 1-3 (v2) - PrivateMath86.093.7+7.7
HardGeoBench <=100miMultimodal36.343.8+7.5
HardGeoBench Country AccuracyMultimodal45.052.5+7.5
GDP.pdf (Artificial Analysis, All-pass)Professional Work24.632.0+7.4
LMArena Agent Arena Steering BurdenAgents-0.513.4+7.0
AutomationBench-AA 1.0.6Agents59.666.6+6.9
Vision Arena OverallMultimodal12661293+6.7
BioSecBench-RespiratoryVirus-Surveillance V1 (110 evaluations)Science5.812.4+6.7
Harvey LAB (Mercor run, undisclosed ~213-task cohort)Legal2.38.8+6.5
Epoch Capabilities Index (ECI)Indexes161169+6.5
GDPval-AA v2.1Professional Work14611607+6.4
Wiz Cyber Model Arena v3.9Cybersecurity64.470.7+6.3
ACCoRDGeneral QA74.780.9+6.2
DAYJOB: HealthcareHealth & Medicine1.67.6+6.0
Finance Agent v2 - Earnings AnalysisFinance & Accounting77.371.5−5.8
MMMU ProMultimodal79.585.1+5.6
MedCodeHealth & Medicine43.448.8+5.4
VCT-Refusal (CAIS)Safety10.716.1+5.4
AA-LCR v1.1Search79.084.0+5.0
MASKSafety93.398.3+5.0
Finance Agent v2 - AdjustmentsFinance & Accounting49.744.7−5.0
HealthBench ProfessionalHealth & Medicine62.467.2+4.8
Machiavelli Deception (CAIS)Safety80.478.8+4.7
Finance Agent v2 - General Qualitative AnalysisFinance & Accounting67.863.2−4.6
Excel Modeling BenchmarkFinance & Accounting66.270.8+4.6
APEX-Accounting 1.1Finance & Accounting59.864.3+4.5
Finance Agent v2 - ComparablesFinance & Accounting47.042.6−4.4
AdvancedIFChat & Writing78.182.4+4.3
CVDP (517-problem set)Code50.955.1+4.2
τ³-Banking (Mercor run, bm25_grep retrieval)Agents38.843.0+4.2
FrontierCode 1.1 Extended - Pass RateCode61.865.9+4.1
CharXiv Extended (Mercor held-out)Multimodal81.185.1+4.0
FrontierCode 1.1 Extended - ScoreCode55.859.8+4.0
Finance Agent v2 - Market AnalysisFinance & Accounting60.664.5+3.8
SWE-Bench Pro V2 — FullCode92.496.1+3.7
MedScribeHealth & Medicine82.986.5+3.6
HealthBench Hard (length-adjusted)Health & Medicine32.736.2+3.5
SAGE v1.1Math87.090.5+3.5
DTBenchGeneral QA93.396.5+3.2
SWE Atlas - Codebase QnACode45.248.4+3.2
Public Benefits Bench v1.1Professional Work62.459.3−3.0
MMLU ProGeneral QA88.191.0+2.9
Tax Agent BenchFinance & Accounting65.262.3−2.9
Legal Research BenchLegal41.338.5−2.9
Finance Agent v2 - All PassFinance & Accounting42.839.9−2.9
Terminal-Bench 2.1Code88.085.4−2.6
ProgramBenchCode0.53.0+2.5
LMArena Text Arena EloChat & Writing14681482+2.5
Financial Coaching BenchFinance & Accounting71.068.5−2.4
DistressBenchSafety76.474.2−2.3
Medical Long Context Reasoning (MLCR-AA)Health & Medicine31.733.9+2.2
LMArena Agent Arena Bash ReliabilityAgents0.75.0+2.1
ObviousBenchGeneral QA97.9100.0+2.1
HealthBench (updated implementation)Health & Medicine58.756.7−2.0
ARC-AGI-1 Semi-PrivateGeneral QA96.598.5+2.0
CharXiv (descriptive and reasoning combined)Multimodal94.196.0+1.9
Finance Agent v2 - General Quantitative AnalysisFinance & Accounting72.774.5+1.9
LMArena Agent Arena Bash RecoveryAgents0.94.5+1.8
CritPt (avg@5)Science30.031.7+1.7
HealthBench (length-adjusted)Health & Medicine57.058.5+1.5
APEX-Agents 1.1Professional Work71.873.2+1.4
SciCode (main problems)Science15.814.6−1.2
Ruby LLM Benchmarks — Program FixerCode74.075.0+1.0
Finance Agent v2 - PrecedentsFinance & Accounting24.825.8+0.9
HealthBench Hard (updated implementation)Health & Medicine34.333.4−0.9
SciCode 1.0.1 (Artificial Analysis)Code55.055.8+0.8
VulcanBench Frontier v4Code89.188.3−0.7
HealthBench Consensus (updated implementation)Health & Medicine95.295.9+0.7
SpeciEval (Refusal-Aware Retries, Historical)Safety98.197.4−0.7
SpeciEval (Pooled English Runs)Safety98.197.4−0.7
SAGEMath47.046.5−0.5
RCM-Bench-1Health & Medicine87.087.4+0.4
OTIS Mock AIME 2024-2025Math99.7100.0+0.3
Finance Agent v2 - Financial ModelingFinance & Accounting23.023.1+0.1
LMArena Agent Arena Tool HallucinationAgents1.01.00.0
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 - 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 - Structured Finance SecuritizationLegal20.020.00.0
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

GPT-5.6 TerravsGPT-6.1 Solnormalised score marginsData Nov 15, 2026Data: Vals AI, Mercor, Arena and 44 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…