Atlas

Compare Models

Claude Fable 5.1 vs GPT-5.5

Claude Fable 5.1 leads by 5.7 capability points (≈95% interval +3.5 to +8.0): a clear lead over GPT-5.5 · P(higher) ≈ 100%

Wins 138 of 166 shared benchmarks (23 lost, 5 tied)

Capability vs. relative cost
Claude Fable 5.1GPT-5.5Closed 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-5.5capability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins138 – 23 · 166 shared, 5 tied

Comparing matching setups only (66 benchmarks excluded) · suites counted once
← Claude Fable 5.1GPT-5.5 →

Claude Fable 5.1vsGPT-5.5wins by categoryData Nov 15, 2026Data: Mercor, Vals AI, Center for AI Safety and 51 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is GPT-5.5 minus Claude Fable 5.1.

VCT-Refusal (CAIS)Safety98.916.0−82.9
HPCT Refusal (CAIS)Safety100.029.0−71.0
MirrorCode (ML, +Private, 2L)Code73.310.0−63.3
Time Horizon Index: KSPAgents63.38.3−55.0
TasteVal (compute multiplier)Science1.70.5−54.7
Medical Long Context Reasoning (MLCR-AA)Health & Medicine71.117.8−53.3
KernelBench Mega Normalized (RTX PRO 6000)Code64.712.2−52.5
GSO Opt@1Code88.240.2−48.0
ChessBench CoherenceGames61.9100.0+38.1
GameBench 2Games94.059.9−34.1
LMArena Agent Arena Net ImprovementAgents14.14.2−32.4
BullshitBench v2 LeaderboardSafety77.047.0−30.0
AA-Briefcase EloAgents16501136−29.9
LLM Debate BenchmarkChat & Writing17381528−29.9
Olam Social Poker EloGames14941521+29.3
BrokenArXiv 05/2026Math78.050.0−28.0
Document ArenaProfessional Work15151484−27.3
Drone-BenchAgents91.263.9−27.3
APEX-SWECode63.636.6−27.0
Code2Video-AA EloCode13251000−26.4
ALE-BenchCode28492108−26.2
BioTIER RefusalSafety99.573.4−26.1
EnterpriseBench: CoreCraftProfessional Work77.451.3−26.1
Furniture Assembly BenchmarkMultimodal70.044.2−25.8
SM BenchSafety89.065.4−23.6
SWE-bench Verified Extended (Mercor held-out)Code61.339.3−22.0
HiL-Bench (Human-in-Loop Benchmark)Agents61.539.7−21.8
SWE Atlas Codebase QnA Extended (Mercor held-out)Code38.017.3−20.7
Humanity's Last ExamGeneral QA60.941.4−19.5
TextQuests Average Cumulative HarmSafety27.121.4+19.4
LiveBench 2.0Agents88.068.7−19.3
Antidote: Everyday EditionSafety10651013−18.6
Terminal-Bench 2.1 Extended (Mercor held-out)Code19.538.0+18.5
GeneBench-Pro Public Release SubsetScience26.08.0−18.0
TaxCalcBench TY25Finance & Accounting42.024.0−18.0
SWE-bench VerifiedCode96.678.8−17.8
SimpleBenchGeneral QA86.669.0−17.6
Tax Agent BenchFinance & Accounting77.960.5−17.4
HANDBOOK.mdProfessional Work38.821.5−17.3
BrokenArXiv 06/2026Math86.669.4−17.1
GDPval-AA v2.1Professional Work17411357−16.9
Hemingway-benchChat & Writing10951035−16.6
LisanBench Path LengthGames151009845−16.3
HardGeoBench Country AccuracyMultimodal61.345.0−16.3
Creative Writing v3 EloChat & Writing21551845−15.7
FrontierMath Tier 4 (v2) - PrivateMath87.872.5−15.3
NoCoT-Bench NCRI 15.2General QA128100−15.3
ChartographyMultimodal46.231.0−15.2
ChessBench EloGames10951349+15.1
Chess Puzzles (task version 2.0)Games48.063.0+15.0
LLM Creative Story WritingChat & Writing3.82.4−15.0
Artificial Analysis Intelligence Index v4.3Indexes53.438.6−14.7
GDPval Extended (Mercor held-out rubric score)Professional Work68.754.5−14.2
GDPval-AA v2Professional Work16501372−13.7
LisanBench Difficulty-WeightedGames45733316−13.5
SWE-bench MultilingualCode91.177.7−13.4
ClockBench AccuracyMultimodal59.446.1−13.3
APEX-AccountingFinance & Accounting61.048.2−12.8
HardGeoBench <=100miMultimodal52.540.0−12.5
FrontierFinance v3.1Finance & Accounting55.943.5−12.4
Design Arena Overall Code CategoriesCode13311257−12.3
NoCoT-Bench NCKI (kspine v3)General QA128111−12.2
Excel Modeling BenchmarkFinance & Accounting76.764.5−12.1
SciCode (main problems)Science24.212.3−11.9
InteractiveBench (task pass)Agents41.730.0−11.7
Clocktower BenchGames15671486−11.6
Finance Agent v2 - General Qualitative AnalysisFinance & Accounting77.065.6−11.4
Finance Agent v2 - Disclosure AnalysisFinance & Accounting70.358.9−11.4
HardGeoBench <=10miMultimodal40.028.8−11.3
LMCAGeneral QA65.554.3−11.2
SpeciEval (Pooled English Runs)Safety87.298.2+11.0
SpeciEval (Refusal-Aware Retries, Historical)Safety87.298.2+11.0
Conceptual Reasoning IndexIndexes74.967.3−10.5
RiemannBenchMath65.655.2−10.4
Finance Agent v2 - PrecedentsFinance & Accounting32.221.9−10.3
BuseyBench SVGMultimodal7.06.0−10.3
ProgramBench (Mercor 100-task subset)Code11.31.0−10.3
ArXivMath 05/2026Math87.577.5−10.0
DeepResearch Bench II (Mercor 130-task set)Search55.045.4−9.6
HardGeoBench Avg DistanceMultimodal8881196−9.5
Finance Agent v2 - Market AnalysisFinance & Accounting69.560.3−9.3
ArXivMath 04/2026Math77.568.8−8.7
Roboflow Vision Evals (October 2026, five tasks)Multimodal74.766.0−8.7
Finance Agent v2 - AdjustmentsFinance & Accounting54.445.7−8.7
Toloka Arena (v1)Agents60.752.3−8.4
Finance Agent v2 - All PassFinance & Accounting47.839.6−8.2
ChessBench AccuracyGames71.378.8+8.0
WeirdML (v2)Science92.984.9−8.0
SUPERChemScience74.766.8−7.9
IntPhys 2Multimodal67.759.9−7.8
SimpleQA Verified (Epoch anti-abstention)General QA70.863.0−7.8
TextQuests Game ProgressGames56.648.9−7.7
AA-AnalystAgent (pass^5)Professional Work57.550.0−7.5
ArXivMath 06/2026Math91.083.6−7.3
Agent Red Teaming k=100 (CAIS)Safety34.241.5−7.3
Longform Creative WritingChat & Writing85.378.2−7.1
GBENCH One-shot Coding GScoreGames69.362.2−7.1
EnigmaEvalGeneral QA44.237.2−7.0
MedXpertQA MM (200-question set)Health & Medicine83.776.7−7.0
SciCode 1.0.1 (Artificial Analysis)Code63.156.1−6.9
APEX-Agents 1.1Professional Work77.170.3−6.8
PCT Political Inconsistency (CAIS)Safety33.440.0−6.6
Roboflow Vision Evals (September 2026 protocol)Multimodal81.374.8−6.5
Extended NYT Connections 940 (Quadratic v1)Games90.096.2+6.2
Finance Agent v2 - General Quantitative AnalysisFinance & Accounting79.773.7−6.0
AutomationBench 1.0.6 (Zapier)Agents22.416.6−5.8
CritPt (avg@5)Science31.125.4−5.7
Blueprint-Bench 2Multimodal0.40.4−5.7
Epoch Capabilities Index (ECI)Indexes167160−5.6
APEX-1 (Mercor, mean score)Professional Work73.868.4−5.4
LiveCodeBench (Vals AI public dataset)Code90.585.3−5.2
MedXpertQA MM Extended (Mercor held-out)Health & Medicine66.060.8−5.2
Senior SWE-BenchCode34.729.5−5.2
ARC-AGI-2General QA90.085.0−5.0
MineBench — Bradley-Terry RatingGames19281846−5.0
MMMU ProMultimodal86.181.2−4.9
FrontierMath Tiers 1-3 (v2) - PrivateMath90.285.3−4.9
ActiveVisionMultimodal15.310.6−4.7
MindCube TinyMultimodal84.579.9−4.6
MedScribeHealth & Medicine91.386.9−4.4
MedCodeHealth & Medicine53.549.1−4.4
ComplexConstraintsChat & Writing45.149.5+4.4
MASKSafety85.990.1+4.2
MMLU ProGeneral QA93.088.9−4.1
MMMU-Pro Extended (Mercor held-out)Multimodal47.643.8−3.8
SpatialViz-BenchMultimodal77.974.2−3.7
LMArena Text Arena EloChat & Writing15031481−3.7
ITBench-AAAgents49.545.8−3.7
GlobeBenchGeneral QA21.625.2+3.6
DAXBench v2Code91.588.3−3.2
AA-LCR v1.1Search87.584.3−3.2
ACCoRDGeneral QA83.780.5−3.2
SAGEMath48.551.5+3.0
DeepSWE v1.1Code67.364.4−2.9
AdvancedIFChat & Writing84.281.3−2.9
Terminal-Bench 2.1Code81.684.3+2.7
AA-LCR Extended (Mercor held-out)Search73.070.5−2.5
ARC-AGI-1 Semi-PrivateGeneral QA97.595.0−2.5
HieroglyphBenchMultimodal20.422.8+2.4
Finance Agent v2 - Financial ModelingFinance & Accounting25.322.9−2.4
Finance Agent v2 - ComparablesFinance & Accounting46.043.7−2.3
Finance Agent v2 - Earnings AnalysisFinance & Accounting75.573.2−2.3
Machiavelli Deception (CAIS)Safety82.481.7+2.1
MortgageTax (Vals all 1,258 documents)Finance & Accounting70.868.8−2.0
GDPval (Mercor 213-task rubric score)Professional Work86.684.6−2.0
Mystery Game PuzzlesGames58.056.0−2.0
LegalBenchLegal88.586.5−2.0
SWE Atlas - Codebase QnACode43.345.2+1.9
Creative Writing v3 Rubric ScoreChat & Writing16.315.9−1.9
Ruby LLM Benchmarks — Program FixerCode76.077.6+1.6
DTBenchGeneral QA97.696.0−1.6
GPQA DiamondGeneral QA93.995.5+1.6
Vision Arena OverallMultimodal12891283−1.5
LMArena Agent Arena Bash RecoveryAgents11.69.5−1.0
TaxEval v2Finance & Accounting76.075.0−1.0
LMArena Agent Arena SteerabilityAgents8.46.9−0.8
ObviousBenchGeneral QA99.3100.0+0.7
BrowseComp (Mercor 130-question subset)Search78.377.7−0.6
CharXiv (descriptive and reasoning combined)Multimodal95.194.6−0.5
CharXiv Extended (Mercor held-out)Multimodal82.282.7+0.5
ERQAMultimodal70.370.5+0.2
AIME 2025Math100.0100.00.0
BabaIsBench Stage 0: The IntroGames100.0100.00.0
CadQueryEval (Corrected Geometry)Code76.076.00.0
FrontierMath ErdősMath0.00.00.0
OTIS Mock AIME 2024-2025Math100.0100.00.0

Claude Fable 5.1vsGPT-5.5normalised score marginsData Nov 15, 2026Data: Mercor, Vals AI, Center for AI Safety and 51 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…