Atlas

Compare Models

Claude Fable 5.1 vs GPT-6.2 Sol

GPT-6.2 Sol leads by 1.4 capability points (≈95% interval −1.2 to +4.0): statistically tied · P(higher) ≈ 85%

Claude Fable 5.1 wins 33 of 54 shared benchmarks (19 lost, 2 tied)

Capability vs. relative cost
Claude Fable 5.1GPT-6.2 SolClosed weightsOpen weights

Benchmark wins33 – 19 · 54 shared, 2 tied

Comparing matching setups only (21 benchmarks excluded) · suites counted once
← Claude Fable 5.1GPT-6.2 Sol →
Data: Artificial Analysis, Surge AI, Arena and 14 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 5.1.

AA-PersonalityBench Informativeness indexLanguage & writing › Style & personality96.237.7−59.1
AA-PersonalityBench Expressiveness indexLanguage & writing › Style & personality98.155.3−43.2
AA-PersonalityBench Transparency indexLanguage & writing › Style & personality98.156.9−41.6
EnterpriseBench: CoreCraftTool & computer use › Customer service & operations77.436.4−41.0
AA-PersonalityBench Warmth indexLanguage & writing › Style & personality94.360.1−34.6
ChartographyPerception & media › Documents & charts46.275.1+28.9
AA-PersonalityBench Imagination indexLanguage & writing › Style & personality86.760.1−26.9
AA-PersonalityBench Conviction indexLanguage & writing › Style & personality10079.2−21.0
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & medicine18.337.2+18.9
AutomationBench 1.0.6 (Zapier)Tool & computer use › Customer service & operations22.441.3+18.9
SimpleBenchKnowledge & reasoning › General knowledge86.669.9−16.7
AA-3D v1.0 EloSoftware › Frontend & design11391259+16.4
SWE Atlas v1.1 - OverallSoftware › Software engineering76.560.5−16.0
Terminal-Bench-Science 0.1Science › Multidisciplinary52.668.0+15.4
Blueprint-Bench 2Perception & media › Spatial & visual reasoning0.40.6+15.1
Tax Agent BenchProfessional › Finance, accounting & tax77.962.9−15.0
AA-PersonalityBench Organization indexLanguage & writing › Style & personality60.045.7−14.5
DAYJOB: Healthcare (October 2026 rescoring)Health & medicine31.217.2−14.0
HANDBOOK.md Agents v1.1Tool & computer use › Customer service & operations47.134.5−12.6
MysteryMechanismMath47.736.9−10.8
DAYJOB: Law (November 2026 release)Professional › Legal22.532.7+10.2
TaxCalcBench TY25Professional › Finance, accounting & tax42.032.0−10.0
FrontierCode 2.0 (Artificial Analysis)Software › Software engineering30.340.2+9.9
ComplexConstraints (November 2026 rescoring)Language & writing › Instruction following60.469.5+9.1
RiemannBench (November 2026 rescoring)Math71.280.0+8.8
Ramp SWE-Bench v2 (Bash harness)Software › Software engineering63.872.4+8.6
AA-PersonalityBench Slop indexLanguage & writing › Style & personality23.815.4−8.6
Professional Reasoning Benchmark - LegalProfessional › Legal51.643.3−8.3
HealthBench Professional (Length-Adjusted)Health & medicine58.165.9+7.8
Humanity's Last Exam DiamondKnowledge & reasoning › General knowledge51.356.5+5.2
DAYJOB: TaxProfessional › Finance, accounting & tax9.64.8−4.8
Harvey LAB-AA v1.1 (hallucination-gated)Professional › Legal6.411.1+4.7
LMArena Agent Arena Net Improvement (Steering Burden)Tool & computer use › General agents13.711.3−4.6
ARC-AGI-2Knowledge & reasoning › Abstract reasoning90.085.8−4.2
sudo L7 (October 2026 rescoring)Software › Software engineering32.236.1+3.9
Dual-Arm Sim (task completion)Tool & computer use › Robotics0.03.8+3.8
AA-AnalystAgent (pass^5)Professional › General knowledge work57.553.8−3.7
DAYJOB: Finance (October 2026 rescoring)Professional › Finance, accounting & tax36.532.8−3.7
LiveBench 2.0 (2026-11-06 release)Composite indexes84.080.3−3.6
AA-Briefcase v1.1 EloProfessional › General knowledge work16711612−3.2
LMArena Agent Arena Bash ReliabilityTool & computer use › General agents9.34.3−2.5
Excel Modeling Benchmark v1.1Professional › Finance, accounting & tax76.774.3−2.4
LMArena Agent Arena Praise vs ComplaintTool & computer use › General agents28.123.3−2.4
GPQA DiamondScience › Multidisciplinary93.796.0+2.3
LMArena Agent Arena Confirmed SuccessTool & computer use › General agents14.210.9−1.6
Gray Swan IPI Attack Success (k=15)Safety & alignment1.02.3−1.3
DualEntry Accounting AI BenchmarkProfessional › Finance, accounting & tax85.184.2−0.9
GDP.xlsxProfessional › General knowledge work23.122.3−0.8
GDP.pdf (November 2026 rescoring)Professional › General knowledge work49.448.8−0.6
DAYJOB: AccountingProfessional › Finance, accounting & tax27.227.6+0.4
LMArena Agent Arena Tool HallucinationTool & computer use › Tool calling1.01.2+0.1
LMArena Agent Arena Steering BurdenTool & computer use › General agents13.513.4−0.1
ARC-AGI-1 Semi-PrivateKnowledge & reasoning › Abstract reasoning97.597.50.0
CritPt v2 (agentic, avg@5)Science › Physical sciences77.777.70.0
Data: Artificial Analysis, Surge AI, Arena and 14 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…