Atlas

Compare Models

Claude Fable 5.1 vs GPT-6.1 Terra

Claude Fable 5.1 leads by 0.4 capability points (≈95% interval −2.1 to +2.9): statistically tied · P(higher) ≈ 63%

Wins 38 of 71 shared benchmarks (31 lost, 2 tied)

Capability vs. relative cost
Claude Fable 5.1GPT-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 Fable 5.1vsGPT-6.1 Terracapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins38 – 31 · 71 shared, 2 tied

Comparing matching setups only (18 benchmarks excluded) · suites counted once
← Claude Fable 5.1GPT-6.1 Terra →

Claude Fable 5.1vsGPT-6.1 Terrawins by categoryData Nov 15, 2026Data: Artificial Analysis, Epoch AI, Mercor and 26 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 Fable 5.1.

BioSecBench-Refusal V1 Clean RerunsSafety5.634.2+28.6
Visa Agentic Commerce BenchmarkAgents75.651.1−24.4
ChessBench CoherenceGames61.984.8+22.9
Tax Agent BenchFinance & Accounting77.955.0−22.8
EBR-bench v4 (card ban, single agent)Games57.178.1+21.0
SpireClimbGames2.022.3+20.3
BioSecBench-Structure V1 (58 evaluations)Science6.325.9+19.5
ChessBenchGames45.364.0+18.7
TaxCalcBench TY25Finance & Accounting42.024.0−18.0
ChartographyMultimodal46.263.2+17.0
τ³-Banking (Mercor run, bm25_grep retrieval)Agents54.037.1−16.9
SWE Atlas v1.1 - OverallCode74.157.7−16.4
Mystery Game PuzzlesGames58.074.0+16.0
Toloka Arena (v3)Agents71.055.0−16.0
Humanity's Last Exam DiamondGeneral QA51.335.7−15.6
AA-Briefcase v1.1 EloAgents16711401−14.8
SimpleQA Verified (Epoch anti-abstention)General QA70.857.3−13.5
FrontierCode 1.1 Main (Artificial Analysis)Code49.036.0−12.9
NoCoT-Bench NCKI (kspine v3)General QA128110−12.9
Harvey LAB (Mercor run, undisclosed ~213-task cohort)Legal17.44.5−12.9
GDPval-AA v2.1Professional Work17651476−12.7
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & Medicine18.330.6+12.2
HardGeoBench <=10miMultimodal45.035.0−10.0
APEX-1 (Mercor, mean score)Professional Work73.864.2−9.6
ChessBench EloGames10951251+9.3
Furniture Assembly BenchmarkMultimodal70.079.2+9.2
Humanity's Last Exam (Text-Only)General QA54.646.2−8.4
LiveBench 2.0Agents88.079.7−8.3
GBENCH One-shot Coding GScoreGames69.377.6+8.3
GameBench 2Games94.085.8−8.2
SciCode 1.0.1 (Artificial Analysis)Code63.155.1−8.0
HardGeoBench Avg DistanceMultimodal8451091−7.6
Terminal-Bench 4.0Code57.950.3−7.6
BioSecBench-RespiratoryVirus-Surveillance V1 (110 evaluations)Science3.611.2+7.6
ChessBench AccuracyGames71.378.3+7.5
HardGeoBench <=100miMultimodal53.846.3−7.5
FrontierCode 2.0 (Artificial Analysis)Code30.337.5+7.2
DualEntry Accounting AI BenchmarkFinance & Accounting85.178.2−6.9
HiL-Bench (Human-in-Loop Benchmark)Agents61.568.0+6.5
LMArena Agent Arena Praise vs ComplaintAgents31.620.0−5.8
Terminal-Bench-Science 0.1Science52.658.3+5.7
DistressBenchSafety71.565.9−5.7
LMArena Agent Arena Net Improvement (Steering Burden)Agents14.812.0−5.3
Design Arena Overall Code CategoriesCode13311300−5.1
TxBench-Antibody-Development V1 (60 evaluations)Science27.822.8−5.0
FrontierMath Tier 4 (v2) - PrivateMath87.892.7+4.9
AA-LCR v1.1Search87.582.7−4.8
HealthBench Professional (Length-Adjusted)Health & Medicine58.162.7+4.6
FORTRESS Average Risk Score (ARS)Safety13.718.2−4.5
Excel Modeling Benchmark v1.1Finance & Accounting76.773.3−3.4
CritPt (avg@5)Science31.128.0−3.1
APEX-Accounting 1.1Finance & Accounting66.763.7−3.0
BuseyBench SVGMultimodal7.07.3+2.5
LMArena Agent Arena Confirmed SuccessAgents14.49.7−2.4
APEX-Agents 1.1Professional Work77.179.3+2.2
DAXBench v2Code91.593.6+2.1
LMArena Agent Arena Bash ReliabilityAgents10.113.4+1.7
MMMU ProMultimodal86.184.7−1.4
NoCoT-Bench NCRI 15.2General QA128130+1.3
ALE-BenchCode28492875+0.9
LMArena Agent Arena Steering BurdenAgents14.816.6+0.9
Epoch Capabilities Index (ECI)Indexes167168+0.9
CritPt v2 (agentic, avg@5)Science77.776.9−0.9
FrontierMath Tiers 1-3 (v2) - PrivateMath90.290.9+0.7
Prophet Arena Arena Score (Agentic, OpenCode)Forecasting86.085.7−0.3
Blueprint-Bench 2Multimodal0.40.4+0.2
Prophet Arena Arena Score (Fixed Context)Forecasting85.085.2+0.2
LMArena Agent Arena Tool HallucinationAgents0.81.0+0.1
GDPxProfessional Work8.58.50.0
HardGeoBench Country AccuracyMultimodal58.858.80.0
OTIS Mock AIME 2024-2025Math100.0100.00.0

Claude Fable 5.1vsGPT-6.1 Terranormalised score marginsData Nov 15, 2026Data: Artificial Analysis, Epoch AI, Mercor and 26 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…