Atlas

Compare Models

Claude Opus 5 vs GPT-6.1 Terra

GPT-6.1 Terra leads by 1.2 capability points (≈95% interval −1.2 to +3.7): statistically tied · P(higher) ≈ 84%

Wins 50 of 85 shared benchmarks (33 lost, 2 tied)

Capability vs. relative cost
Claude Opus 5GPT-6.1 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 Opus 5vsGPT-6.1 Terracapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins33 – 50 · 85 shared, 2 tied

Comparing matching setups only (10 benchmarks excluded) · suites counted once
← Claude Opus 5GPT-6.1 Terra →

Claude Opus 5vsGPT-6.1 Terrawins by categoryData Nov 15, 2026Data: Artificial Analysis, benchmarks.bio, Epoch AI and 23 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is GPT-6.1 Terra minus Claude Opus 5.

EBR-bench v4 (card ban, single agent)Games45.778.1+32.4
Legal Research BenchLegal55.326.0−29.3
Terminal-Bench-Science 0.1Science28.657.6+29.0
ExploitBench - release-reported exploit progressCybersecurity70.097.1+27.1
ChessBench EloGames8261251+25.3
OncologyBench Strict (25 tasks, October 9 rubric)Health & Medicine40.016.0−24.0
Tax Agent BenchFinance & Accounting76.755.0−21.6
ChessBench CoherenceGames63.884.8+21.0
FrontierMath Tier 4 (v2) - PrivateMath73.292.7+19.5
Furniture Assembly BenchmarkMultimodal60.879.2+18.3
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & Medicine13.930.6+16.7
VariantBench V1 Clean RerunsScience50.634.8−15.8
MetagenomicsBench V1Science54.038.7−15.3
Mystery Game PuzzlesGames59.074.0+15.0
AA-Briefcase v1.1 EloAgents16541401−13.9
Vending-Bench 2Games111829107−12.6
SpatialBench-Long V1 Clean RerunsScience28.040.5+12.5
MacroscopeBenchCode69.982.3+12.4
TxBench Oligo Discovery V1 (113 evaluations)Science42.530.4−12.1
Blueprint-Bench 2Multimodal0.30.4+11.7
BioSecBench-Surveillance V1 Clean RerunsScience51.539.9−11.6
FrontierCode 2.0 (Artificial Analysis)Code31.242.3+11.2
GDPval-AA v2.1Professional Work17291476−11.2
HiL-Bench (Human-in-Loop Benchmark)Agents57.068.0+11.0
BioSecBench-Structure V1 (58 evaluations)Science15.525.9+10.3
CritPt v2 (agentic, avg@5)Science71.481.4+10.0
BuseyBench SVGMultimodal6.37.3+9.8
BioSecBench-Refusal V1 Clean RerunsSafety24.534.2+9.7
GBENCH One-shot Coding GScoreGames68.177.6+9.4
LMArena Agent Arena Net Improvement (Steering Burden)Agents7.212.0+9.0
ALE-BenchCode26222875+8.9
ScBench-Long V1 Clean RerunsScience27.318.8−8.5
NoCoT-Bench NCRI 15.2General QA115130+8.4
Humanity's Last Exam (Text-Only)General QA54.446.2−8.2
TxBench Antibody Discovery V1 (100 evaluations)Science48.740.7−8.0
ChessBench AccuracyGames71.078.3+7.8
DAXBench v2Code86.293.6+7.4
TxBench Preclinical Pharmacology V1 Clean RerunsScience63.055.7−7.3
Artificial Analysis Intelligence Index v5.0.1Indexes44.251.3+7.2
GDP.pdf (Artificial Analysis, All-pass)Professional Work21.628.6+7.0
Toloka Arena (v3)Agents62.055.0−7.0
τ³-Banking (Mercor run, bm25_grep retrieval)Agents44.037.1−6.9
NoCoT-Bench NCKI (kspine v3)General QA119110−6.7
Terminal-Bench 4.0 (Artificial Analysis)Code49.055.6+6.6
GPQA DiamondGeneral QA87.994.3+6.4
LMArena Agent Arena Steering BurdenAgents4.416.6+6.1
FrontierMath Tiers 1-3 (v2) - PrivateMath85.690.9+5.3
HardGeoBench Country AccuracyMultimodal53.858.8+5.0
SpatialBench V1 Clean RerunsScience68.473.3+4.9
Vibe Code Bench v1.1Code88.484.0−4.4
APEX-1 (Mercor, mean score)Professional Work68.664.2−4.4
LLM Chess EloGames12851391+4.3
DistressBenchSafety70.265.9−4.3
Agents' Last Exam Overall ScoreAgents55.960.0+4.1
Artificial Analysis Progress Index v1Indexes334345+4.0
TaxCalcBench TY25Finance & Accounting20.024.0+4.0
LMArena Agent Arena Praise vs ComplaintAgents12.220.0+3.9
HardGeoBench Avg DistanceMultimodal12181091+3.9
AA-Omniscience IndexIndexes37.129.8−3.6
Terminal-Bench 4.0Code53.950.3−3.6
AutomationBench-AA 1.0.6Agents56.660.2+3.6
Design Arena Overall Code CategoriesCode13211300−3.5
BioSecBench-Function V1 Clean RerunsScience42.539.0−3.4
BioSecBench-RespiratoryVirus-Surveillance V1 (110 evaluations)Science14.611.2−3.3
Epoch Capabilities Index (ECI)Indexes164168+2.9
Artificial Analysis Intelligence Index v4.3.2Indexes50.848.1−2.7
HardGeoBench <=10miMultimodal32.535.0+2.5
Humanity's Last Exam DiamondGeneral QA44.847.3+2.5
LiveBench 2.0Agents82.179.7−2.4
APEX-Agents 1.1Professional Work77.079.3+2.3
EpiBench V1 Clean RerunsScience25.223.3−1.9
LMArena Agent Arena Bash ReliabilityAgents9.813.4+1.8
FORTRESS Average Risk Score (ARS)Safety16.818.2−1.3
SciCode 1.0.1 (Artificial Analysis)Code56.455.1−1.3
CritPt (avg@5)Science29.128.0−1.1
OTIS Mock AIME 2024-2025Math98.9100.0+1.1
GameBench 2Games84.885.8+1.0
VoxelBench Text RatingGames23002313+0.9
MMMU ProMultimodal84.084.7+0.7
AA-LCR v1.1Search82.082.7+0.7
APEX-Accounting 1.1Finance & Accounting64.363.7−0.6
LMArena Agent Arena Confirmed SuccessAgents10.29.7−0.3
LMArena Agent Arena Tool HallucinationAgents0.91.00.0
HardGeoBench <=100miMultimodal46.346.30.0
ScBench V1 Clean RerunsScience58.358.30.0

Claude Opus 5vsGPT-6.1 Terranormalised score marginsData Nov 15, 2026Data: Artificial Analysis, benchmarks.bio, Epoch AI and 23 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…