Atlas

Compare Models

Claude Fable 5.2 vs GPT-6.1 Sol

Claude Fable 5.2 leads by 4.0 capability points (≈95% interval +1.1 to +6.8): a clear lead over GPT-6.1 Sol · P(higher) ≈ 100%

Wins 65 of 83 shared benchmarks (14 lost, 4 tied)

Capability vs. relative cost
Claude Fable 5.2GPT-6.1 SolClosed 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 5.2vsGPT-6.1 Solcapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins65 – 14 · 83 shared, 4 tied

Comparing matching setups only (16 benchmarks excluded) · suites counted once
← Claude Fable 5.2GPT-6.1 Sol →

Claude Fable 5.2vsGPT-6.1 Solwins by categoryData Nov 15, 2026Data: Artificial Analysis, Vals AI, Epoch AI and 22 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is GPT-6.1 Sol minus Claude Fable 5.2.

HoneyBench v0.2 (hack rate)Safety11.959.5−47.6
EBR-bench v4 (card ban, single agent)Games90.554.3−36.2
LMArena Agent Arena Net Improvement (Steering Burden)Agents30.412.5−33.8
AA-PersonalityBench Expressiveness indexChat & Writing81.050.5−30.8
AA-PersonalityBench Organization indexChat & Writing54.381.0+26.9
LMArena Agent Arena Praise vs ComplaintAgents73.427.5−23.0
Humanity's Last ExamGeneral QA73.251.3−21.9
AA-3D v1.0 EloCode12711121−21.8
Humanity's Last Exam DiamondGeneral QA80.258.8−21.4
HardGeoBench Avg DistanceMultimodal5521239−21.2
AA-PersonalityBench Conviction indexChat & Writing75.354.3−21.2
FrontierCode 1.1 Main (Artificial Analysis)Code64.043.3−20.7
BioSecBench-Structure V1 (58 evaluations)Science41.421.3−20.1
SimpleQA Verified (Epoch anti-abstention)General QA94.073.9−20.1
AA-PersonalityBench Informativeness indexChat & Writing46.727.7−19.2
SWE-bench Multimodal (Vals v1, 480 Tasks)Code62.343.5−18.7
HardGeoBench Country AccuracyMultimodal70.052.5−17.5
AA-PersonalityBench Warmth indexChat & Writing82.965.7−17.3
HardGeoBench <=100miMultimodal60.043.8−16.2
APEX-Agents 1.1Professional Work87.673.2−14.4
HardGeoBench <=10miMultimodal48.835.0−13.7
OpenWorldEngineer (Mean@8)Code54.641.2−13.4
τ³-Banking (Mercor run, bm25_grep retrieval)Agents55.743.0−12.7
AA-Briefcase v1.1 EloAgents17851555−12.6
HieroglyphBenchMultimodal43.631.4−12.2
SciCode 1.0.1 (Artificial Analysis)Code67.755.8−11.9
SWE-bench MultilingualCode96.684.9−11.7
AA-PersonalityBench Transparency indexChat & Writing81.092.4+11.5
Chess Puzzles (task version 2.0)Games79.068.0−11.0
Furniture Assembly BenchmarkMultimodal90.080.0−10.0
Mystery Game PuzzlesGames90.080.0−10.0
Conceptual Reasoning IndexIndexes79.972.7−9.9
GDPval-AA v2.1Professional Work18311607−9.9
LMArena Agent Arena Steering BurdenAgents33.013.4−9.8
ACCoRDGeneral QA89.480.9−8.5
FrontierCode 2.0 (Artificial Analysis)Code48.840.6−8.1
GDPxProfessional Work16.48.4−8.0
LMCAGeneral QA70.662.8−7.8
BuseyBench SVGMultimodal7.28.0+7.8
Vals Finance IndexIndexes63.655.8−7.8
AA-PersonalityBench Slop indexChat & Writing6.714.3+7.7
Finance Agent v2.1Finance & Accounting69.561.8−7.7
AutomationBench-AA 1.0.6Agents74.266.6−7.6
GameBench 2Games99.192.0−7.1
Excel Modeling Benchmark v1.1Finance & Accounting77.471.1−6.3
Immigration BenchLegal74.067.8−6.2
APEX-Accounting 1.1Finance & Accounting70.464.3−6.1
Ramp Accounting BenchFinance & Accounting52.246.4−5.8
AA-PersonalityBench Imagination indexChat & Writing67.661.9−5.8
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & Medicine43.938.3−5.6
MysteryMechanismScience51.846.4−5.4
LMArena Agent Arena Confirmed SuccessAgents22.011.7−5.1
Vibe Code Bench v1.1Code93.888.9−4.8
Agent Initiative BenchmarkAgents8.28.7+4.8
LMArena Agent Arena Bash ReliabilityAgents14.45.0−4.7
Tax Agent BenchFinance & Accounting67.062.3−4.7
Epoch Capabilities Index (ECI)Indexes174169−4.3
DAXBench v2Code88.392.6+4.3
SpireClimbGames19.623.8+4.2
BioMysteryBench (Vals Terminus 2)Science83.779.6−4.1
CritPt (avg@5)Science35.731.7−4.0
CritPt v2 (agentic, avg@5)Science86.382.6−3.7
HealthBenchHealth & Medicine60.156.9−3.2
HealthBench ProfessionalHealth & Medicine71.468.3−3.1
CrossSearch (pass^3)Search56.059.0+3.0
DTBenchGeneral QA99.296.5−2.7
HealthBench (length-adjusted, Artificial Analysis)Health & Medicine60.458.5−1.9
FrontierMath Tiers 1-3 (v2) - PrivateMath95.193.7−1.4
MacroscopeBenchCode78.577.2−1.3
HLAB-Vals-1.1Legal5.06.3+1.3
Harvey LAB (Mercor run, undisclosed ~213-task cohort)Legal10.08.8−1.2
ProofBench v1.1Math100.099.0−1.0
DeepSWE v1.1Code74.375.2+0.9
ARC-AGI-3 (Semi-Private)General QA97.396.4−0.9
Clocktower BenchGames17591764+0.7
AA-LCR v1.1Search83.384.0+0.7
ChessBenchGames69.368.7−0.7
ARC-AGI-2General QA93.894.2+0.4
OTIS Mock AIME 2024-2025Math99.7100.0+0.3
LMArena Agent Arena Tool HallucinationAgents1.01.00.0
ARC-AGI-1 Semi-PrivateGeneral QA98.598.50.0
DualEntry Accounting AI BenchmarkFinance & Accounting81.281.20.0
FrontierMath Tier 4 (v2) - PrivateMath100.0100.00.0

Claude Fable 5.2vsGPT-6.1 Solnormalised score marginsData Nov 15, 2026Data: Artificial Analysis, Vals AI, Epoch AI and 22 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…