Atlas

Compare Models

Claude Sonnet 5.5 vs GPT-5.6 Terra

Claude Sonnet 5.5 leads by 5.9 capability points (≈95% interval +3.6 to +8.3): a clear lead over GPT-5.6 Terra · P(higher) ≈ 100%

Wins 121 of 165 shared benchmarks (21 lost, 23 tied)

Capability vs. relative cost
Claude Sonnet 5.5GPT-5.6 TerraClosed 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 Sonnet 5.5vsGPT-5.6 Terracapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins121 – 21 · 165 shared, 23 tied

Comparing matching setups only (53 benchmarks excluded) · suites counted once
← Claude Sonnet 5.5GPT-5.6 Terra →

Claude Sonnet 5.5vsGPT-5.6 Terrawins by categoryData Nov 15, 2026Data: Vals AI, Mercor, Artificial Analysis and 38 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is GPT-5.6 Terra minus Claude Sonnet 5.5.

Medical Long Context Reasoning (MLCR-AA)Health & Medicine75.031.7−43.3
HardGeoBench Avg DistanceMultimodal6831991−40.3
Terminal-Bench-Science 0.1Science53.313.8−39.5
YC-BenchGames1741315636144−38.6
GameBench 2Games91.453.5−38.0
Terminal-Bench Science 0.1 (Vals)Science47.110.0−37.1
Terminal-Bench 4.0Code61.826.3−35.6
APEX-SWECode66.431.1−35.3
LMArena Agent Arena Net Improvement (Steering Burden)Agents16.0-1.8−33.6
LisanBench Path LengthGames137873303−32.6
GeneBench-Pro Public Release SubsetScience39.08.0−31.0
Mystery Game PuzzlesGames65.035.0−30.0
SWE Atlas Codebase QnA Extended (Mercor held-out)Code44.716.0−28.7
Terminal-Bench 4.0 (Artificial Analysis)Code63.635.4−28.3
ChartographyMultimodal61.634.0−27.6
AA-Briefcase v1.1 EloAgents18161333−26.5
SM BenchSafety85.459.0−26.4
ProofBench v1.1Math100.074.0−26.0
Humanity's Last Exam DiamondGeneral QA53.328.8−24.5
GDPval Extended (Mercor held-out rubric score)Professional Work70.947.2−23.7
ARC-AGI-3 (Semi-Private)General QA24.30.7−23.7
Vending-Bench 2Games111657343−23.1
DAYJOB: FinanceFinance & Accounting24.01.8−22.2
Artificial Analysis Intelligence Index v5.0.1Indexes54.532.3−22.2
Code MigrationCode69.847.8−22.0
SWE-bench Verified Extended (Mercor held-out)Code57.378.7+21.4
Furniture Assembly BenchmarkMultimodal75.054.2−20.8
Harvey's Legal Agent Benchmark - Data Privacy CybersecurityLegal20.00.0−20.0
Harvey's Legal Agent Benchmark - Structured Finance SecuritizationLegal0.020.0+20.0
Harvey's Legal Agent Benchmark - Trusts Estates Private ClientLegal20.00.0−20.0
LisanBench Difficulty-WeightedGames29831197−19.1
Concept Synth INDUCTION Challenge100 Round 1General QA44.025.0−19.0
SWE-bench VerifiedCode96.777.8−18.9
DeepsecBench — Revalidated V1 (233 issues)Code37.519.1−18.4
Design Arena Overall Code CategoriesCode13491238−18.4
FrontierCode 2.0 (Artificial Analysis)Code41.723.9−17.8
Vibe Code Bench v1.1Code92.474.6−17.8
GDPval-AA v2.1Professional Work18381461−16.6
Olam Social Poker EloGames15261511−16.3
LMArena Agent Arena Praise vs ComplaintAgents30.5-1.9−16.2
AA-Omniscience IndexIndexes32.30.1−16.1
Conceptual Reasoning IndexIndexes77.165.7−15.7
APEX-1 (Mercor, mean score)Professional Work53.969.5+15.6
τ³-Banking (Mercor run, bm25_grep retrieval)Agents54.038.8−15.2
Bio Capabilities Index (SecureBio)Indexes175155−14.6
CursorBench 4.0Code55.541.3−14.2
Artificial Analysis Intelligence Index v4.3.2Indexes56.042.1−13.9
HardGeoBench <=100miMultimodal50.036.3−13.7
HardGeoBench Country AccuracyMultimodal58.845.0−13.7
DAYJOB: HealthcareHealth & Medicine15.21.6−13.6
LMArena Agent Arena Confirmed SuccessAgents19.5-7.2−13.4
LMCAGeneral QA68.355.0−13.3
Concept Synth INDUCTION Challenge64 Round 1General QA50.037.5−12.5
AutomationBench-AA 1.0.6Agents71.859.6−12.1
Chess Puzzles (task version 2.0)Games59.047.0−12.0
APEX-Agents 1.1Professional Work83.771.8−11.9
Roboflow Vision Evals (October 2026, five tasks)Multimodal77.666.0−11.6
BioSecBench-RespiratoryVirus-Surveillance V1 (110 evaluations)Science17.35.8−11.5
LAB-Bench ProtocolQA (SecureBio)Science88.377.0−11.3
HardGeoBench <=10miMultimodal36.325.0−11.2
ACCoRDGeneral QA85.774.7−11.0
ALE-BenchCode24482154−10.4
DeepSWE v1.1Code70.560.2−10.3
VCT v2 (SecureBio)Science52.241.9−10.3
Humanity's Last ExamGeneral QA46.135.9−10.2
Eval Connections One-Shot with TrapsGames76.086.0+10.0
Roboflow Vision Evals (September 2026 protocol)Multimodal83.873.8−10.0
Harvey HLAB — Criteria Pass RateLegal70.079.8+9.8
FrontierMath Tier 4 (v2) - PrivateMath80.570.7−9.8
CVDP (517-problem set)Code60.650.9−9.7
Excel Modeling BenchmarkFinance & Accounting75.766.2−9.5
MedCodeHealth & Medicine52.943.4−9.5
Artificial Analysis Progress Index v1Indexes351324−9.4
Terminal-Bench 2.1 Extended (Mercor held-out)Code45.536.4−9.1
HPCT (SecureBio)Science77.268.2−9.0
Blueprint-Bench 2Multimodal0.40.3−8.5
MedScribeHealth & Medicine91.182.9−8.2
LMArena Agent Arena Steering BurdenAgents14.5-0.5−7.5
RedlineBenchLegal40.347.8+7.5
LiveBench 2.0Agents82.575.0−7.5
SpeciEval (Refusal-Aware Retries, Historical)Safety91.098.1+7.1
SpeciEval (Pooled English Runs)Safety91.098.1+7.1
LMArena Agent Arena Bash RecoveryAgents15.00.9−7.1
WCB (SecureBio)Science66.559.8−6.7
Quant-Arena (iterative total score)Finance & Accounting40.947.6+6.6
MMMU ProMultimodal85.879.5−6.3
SciCode 1.0.1 (Artificial Analysis)Code61.055.0−6.0
BullshitBench v2 LeaderboardSafety60.054.0−6.0
CritPt v2 (agentic, avg@5)Science75.769.7−6.0
ProgramBenchCode6.50.5−6.0
LMArena Agent Arena Bash ReliabilityAgents12.70.7−6.0
GDPval (Mercor 213-task rubric score)Professional Work86.580.7−5.8
AdvancedIFChat & Writing83.478.1−5.3
VCT (SecureBio)Science52.547.2−5.3
ChessBench CoherenceGames64.669.7+5.1
SWE Atlas - Codebase QnACode50.345.2−5.1
MedXpertQA MM (200-question set)Health & Medicine80.775.7−5.0
Public Benefits Bench v1.1Professional Work67.262.4−4.8
SAGEMath51.847.0−4.8
APEX-Accounting 1.1Finance & Accounting64.559.8−4.7
LAB-Bench LitQA2 (SecureBio)Science55.851.1−4.7
MedXpertQA MM Extended (Mercor held-out)Health & Medicine55.651.0−4.6
IOI (Vals v2)Code83.187.6+4.6
BuseyBench SVGMultimodal6.56.1−4.5
DeepResearch Bench II (Mercor 130-task set)Search51.747.2−4.5
LAB-Bench CloningScenarios (SecureBio)Science97.092.5−4.5
Epoch Capabilities Index (ECI)Indexes166161−4.4
ChessBench AccuracyGames75.871.9−4.2
GBENCH One-shot Coding GScoreGames67.463.3−4.0
ChessBench EloGames857925+4.0
LLM Chess EloGames11951293+4.0
DTBenchGeneral QA97.393.3−4.0
Financial Coaching BenchFinance & Accounting74.971.0−4.0
SWE-bench MultilingualCode90.994.4+3.5
SWE-Bench Pro V2 — FullCode95.892.4−3.4
NoCoT-Bench NCRI 15.2General QA95.189.0−3.3
CharXiv Extended (Mercor held-out)Multimodal84.281.1−3.1
MMLU ProGeneral QA91.088.1−2.9
WMDP-Bio (SecureBio)Science92.689.7−2.9
FrontierMath Tiers 1-3 (v2) - PrivateMath88.886.0−2.8
NoCoT-Bench NCKI (kspine v3)General QA107104−2.3
GPQA-Bio (SecureBio)Science87.685.3−2.3
Extended NYT Connections 940 (Quadratic v1)Games80.578.4−2.1
ObviousBenchGeneral QA100.097.9−2.1
CharXiv (descriptive and reasoning combined)Multimodal96.194.1−2.0
MBCT (SecureBio)Science58.856.9−1.9
DistressBenchSafety74.676.4+1.9
CritPt (avg@5)Science31.430.0−1.4
GDP.pdf (Artificial Analysis, All-pass)Professional Work25.824.6−1.2
SAGE v1.1Math88.087.0−1.0
Vision Arena OverallMultimodal12631266+0.9
PubMedQA (SecureBio)Health & Medicine81.380.4−0.9
LMArena Text Arena EloChat & Writing14721468−0.8
BrowseComp (Mercor 130-question subset)Search86.685.8−0.8
Terminal-Bench 2.1Code87.388.0+0.7
AA-LCR v1.1Search79.779.0−0.7
AA-LCR Extended (Mercor held-out)Search73.874.4+0.6
MMLU-Bio (SecureBio)Science91.090.7−0.3
OTIS Mock AIME 2024-2025Math100.099.7−0.3
Ruby LLM Benchmarks — Program FixerCode73.874.0+0.2
LMArena Agent Arena Tool HallucinationAgents0.91.00.0
ARC-AGI-2General QA83.983.90.0
AIME 2025Math100.0100.00.0
ARC-AGI-1 Semi-PrivateGeneral QA96.596.50.0
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 - Emerging Companies Venture CapitalLegal0.00.00.0
Harvey's Legal Agent Benchmark - Employment LaborLegal0.00.00.0
Harvey's Legal Agent Benchmark - Energy Natural ResourcesLegal0.00.00.0
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

Claude Sonnet 5.5vsGPT-5.6 Terranormalised score marginsData Nov 15, 2026Data: Vals AI, Mercor, Artificial Analysis and 38 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…