Atlas

Compare Models

Claude Fable 5.2 vs GPT-6.2 Sol

Claude Fable 5.2 leads by 2.8 capability points (≈95% interval −0.1 to +5.7): statistically tied · P(higher) ≈ 97%

Wins 47 of 61 shared benchmarks (13 lost, 1 tied)

Capability vs. relative cost
Claude Fable 5.2GPT-6.2 SolClosed weightsOpen weights

Benchmark wins47 – 13 · 61 shared, 1 tied

Comparing matching setups only (37 benchmarks excluded) · suites counted once
← Claude Fable 5.2GPT-6.2 Sol →
Data: Artificial Analysis, Surge AI, Vals AI and 17 othersatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is GPT-6.2 Sol minus Claude Fable 5.2.

EnterpriseBench: CoreCraftTool & computer use › Customer service & operations68.236.4−31.8
LMArena Agent Arena Net Improvement (Steering Burden)Tool & computer use › General agents27.411.3−31.4
KradleBenchGames37.867.9+30.2
TaxCalcBench TY25Professional › Finance, accounting & tax62.032.0−30.0
DAYJOB: Healthcare (October 2026 rescoring)Health & medicine45.017.2−27.8
AA-PersonalityBench Expressiveness indexLanguage & writing › Style & personality81.055.3−25.9
SimpleBenchKnowledge & reasoning › General knowledge94.969.9−25.0
AA-PersonalityBench Transparency indexLanguage & writing › Style & personality81.056.9−24.3
AA-PersonalityBench Warmth indexLanguage & writing › Style & personality82.960.1−23.0
LMArena Agent Arena Praise vs ComplaintTool & computer use › General agents68.723.3−22.7
SWE Atlas v1.1 - OverallSoftware › Software engineering80.560.5−20.1
HANDBOOK.md Agents v1.1Tool & computer use › Customer service & operations52.334.5−17.8
DAYJOB: AccountingProfessional › Finance, accounting & tax45.227.6−17.6
Humanity's Last Exam DiamondKnowledge & reasoning › General knowledge80.264.0−16.2
sudo L7 (October 2026 rescoring)Software › Software engineering52.236.1−16.1
MysteryMechanismMath51.836.9−14.9
DAYJOB: Finance (October 2026 rescoring)Professional › Finance, accounting & tax43.732.8−10.9
DAYJOB: TaxProfessional › Finance, accounting & tax15.64.8−10.8
GDP.xlsxProfessional › General knowledge work32.922.3−10.6
Finance Agent v2.1Professional › Finance, accounting & tax69.558.9−10.6
AA-Briefcase v1.1 EloProfessional › General knowledge work17851612−9.5
AA-PersonalityBench Informativeness indexLanguage & writing › Style & personality46.737.7−9.1
AA-PersonalityBench Slop indexLanguage & writing › Style & personality6.715.4+8.7
AA-PersonalityBench Organization indexLanguage & writing › Style & personality54.345.7−8.7
MultiChallengeLanguage & writing › Conversation80.472.0−8.4
ARC-AGI-2Knowledge & reasoning › Abstract reasoning93.885.8−7.9
CritPt v2 (agentic, avg@5)Science › Physical sciences86.378.6−7.7
AA-PersonalityBench Imagination indexLanguage & writing › Style & personality67.660.1−7.6
HLAB-Vals-1.1Professional › Legal5.012.1+7.1
Vals Finance IndexProfessional › Finance, accounting & tax63.656.7−6.9
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & medicine43.937.2−6.7
DAYJOB: Law (November 2026 release)Professional › Legal26.532.7+6.2
Vibe Code Bench v1.1Software › Frontend & design93.887.6−6.2
BioMysteryBench (Vals Terminus 2)Science › Life sciences83.777.8−5.9
LMArena Agent Arena Confirmed SuccessTool & computer use › General agents22.510.9−5.8
Professional Reasoning Benchmark - LegalProfessional › Legal49.043.3−5.6
LMArena Agent Arena Steering BurdenTool & computer use › General agents24.513.4−5.6
ComplexConstraints (November 2026 rescoring)Language & writing › Instruction following64.569.5+5.0
LiveBench 2.0 (2026-11-06 release)Composite indexes84.980.3−4.6
LMArena Agent Arena Bash ReliabilityTool & computer use › General agents13.14.3−4.4
FrontierCode 2.0 (Artificial Analysis)Software › Software engineering48.844.4−4.4
Tax Agent BenchProfessional › Finance, accounting & tax67.062.9−4.1
AA-PersonalityBench Conviction indexLanguage & writing › Style & personality75.379.2+4.0
Dual-Arm Sim (task completion)Tool & computer use › Robotics7.53.8−3.7
RiemannBench (November 2026 rescoring)Math83.280.0−3.2
Excel Modeling Benchmark v1.1Professional › Finance, accounting & tax77.474.3−3.1
AutomationBench 1.0.6 (Zapier)Tool & computer use › Customer service & operations44.341.3−3.0
GDP.pdf (November 2026 rescoring)Professional › General knowledge work51.848.8−3.0
DualEntry Accounting AI BenchmarkProfessional › Finance, accounting & tax81.284.2+3.0
AutomationBench-AA 1.0.6Tool & computer use › Customer service & operations74.271.3−2.9
Humanity's Sixth SensePerception & media › Spatial & visual reasoning59.556.8−2.7
GDPval v2 (OpenAI internal)Professional › General knowledge work44.846.9+2.1
Gray Swan IPI Attack Success (k=15)Safety & alignment4.12.3+1.8
ChartographyPerception & media › Documents & charts73.475.1+1.7
AA-3D v1.0 EloSoftware › Frontend & design12711259−1.7
ARC-AGI-1 Semi-PrivateKnowledge & reasoning › Abstract reasoning98.597.5−1.0
Ramp SWE-Bench v2 (Bash harness)Software › Software engineering73.072.4−0.6
Blueprint-Bench 2Perception & media › Spatial & visual reasoning0.60.6+0.4
Agents' Last Exam v1.1 ScoreProfessional › General knowledge work72.873.0+0.3
ARC-AGI-3 (Semi-Private)Games97.397.4+0.1
LMArena Agent Arena Tool HallucinationTool & computer use › Tool calling1.21.20.0
Data: Artificial Analysis, Surge AI, Vals AI and 17 othersatlas.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…