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
Area← 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: CoreCraft | Tool & computer use › Customer service & operations | 68.2 | 36.4 | −31.8 |
| LMArena Agent Arena Net Improvement (Steering Burden) | Tool & computer use › General agents | 27.4 | 11.3 | −31.4 |
| KradleBench | Games | 37.8 | 67.9 | +30.2 |
| TaxCalcBench TY25 | Professional › Finance, accounting & tax | 62.0 | 32.0 | −30.0 |
| DAYJOB: Healthcare (October 2026 rescoring) | Health & medicine | 45.0 | 17.2 | −27.8 |
| AA-PersonalityBench Expressiveness index | Language & writing › Style & personality | 81.0 | 55.3 | −25.9 |
| SimpleBench | Knowledge & reasoning › General knowledge | 94.9 | 69.9 | −25.0 |
| AA-PersonalityBench Transparency index | Language & writing › Style & personality | 81.0 | 56.9 | −24.3 |
| AA-PersonalityBench Warmth index | Language & writing › Style & personality | 82.9 | 60.1 | −23.0 |
| LMArena Agent Arena Praise vs Complaint | Tool & computer use › General agents | 68.7 | 23.3 | −22.7 |
| SWE Atlas v1.1 - Overall | Software › Software engineering | 80.5 | 60.5 | −20.1 |
| HANDBOOK.md Agents v1.1 | Tool & computer use › Customer service & operations | 52.3 | 34.5 | −17.8 |
| DAYJOB: Accounting | Professional › Finance, accounting & tax | 45.2 | 27.6 | −17.6 |
| Humanity's Last Exam Diamond | Knowledge & reasoning › General knowledge | 80.2 | 64.0 | −16.2 |
| sudo L7 (October 2026 rescoring) | Software › Software engineering | 52.2 | 36.1 | −16.1 |
| MysteryMechanism | Math | 51.8 | 36.9 | −14.9 |
| DAYJOB: Finance (October 2026 rescoring) | Professional › Finance, accounting & tax | 43.7 | 32.8 | −10.9 |
| DAYJOB: Tax | Professional › Finance, accounting & tax | 15.6 | 4.8 | −10.8 |
| GDP.xlsx | Professional › General knowledge work | 32.9 | 22.3 | −10.6 |
| Finance Agent v2.1 | Professional › Finance, accounting & tax | 69.5 | 58.9 | −10.6 |
| AA-Briefcase v1.1 Elo | Professional › General knowledge work | 1785 | 1612 | −9.5 |
| AA-PersonalityBench Informativeness index | Language & writing › Style & personality | 46.7 | 37.7 | −9.1 |
| AA-PersonalityBench Slop index | Language & writing › Style & personality | 6.7 | 15.4 | +8.7 |
| AA-PersonalityBench Organization index | Language & writing › Style & personality | 54.3 | 45.7 | −8.7 |
| MultiChallenge | Language & writing › Conversation | 80.4 | 72.0 | −8.4 |
| ARC-AGI-2 | Knowledge & reasoning › Abstract reasoning | 93.8 | 85.8 | −7.9 |
| CritPt v2 (agentic, avg@5) | Science › Physical sciences | 86.3 | 78.6 | −7.7 |
| AA-PersonalityBench Imagination index | Language & writing › Style & personality | 67.6 | 60.1 | −7.6 |
| HLAB-Vals-1.1 | Professional › Legal | 5.0 | 12.1 | +7.1 |
| Vals Finance Index | Professional › Finance, accounting & tax | 63.6 | 56.7 | −6.9 |
| Medical Long Context Reasoning (MLCR-AA, October 2026 methodology) | Health & medicine | 43.9 | 37.2 | −6.7 |
| DAYJOB: Law (November 2026 release) | Professional › Legal | 26.5 | 32.7 | +6.2 |
| Vibe Code Bench v1.1 | Software › Frontend & design | 93.8 | 87.6 | −6.2 |
| BioMysteryBench (Vals Terminus 2) | Science › Life sciences | 83.7 | 77.8 | −5.9 |
| LMArena Agent Arena Confirmed Success | Tool & computer use › General agents | 22.5 | 10.9 | −5.8 |
| Professional Reasoning Benchmark - Legal | Professional › Legal | 49.0 | 43.3 | −5.6 |
| LMArena Agent Arena Steering Burden | Tool & computer use › General agents | 24.5 | 13.4 | −5.6 |
| ComplexConstraints (November 2026 rescoring) | Language & writing › Instruction following | 64.5 | 69.5 | +5.0 |
| LiveBench 2.0 (2026-11-06 release) | Composite indexes | 84.9 | 80.3 | −4.6 |
| LMArena Agent Arena Bash Reliability | Tool & computer use › General agents | 13.1 | 4.3 | −4.4 |
| FrontierCode 2.0 (Artificial Analysis) | Software › Software engineering | 48.8 | 44.4 | −4.4 |
| Tax Agent Bench | Professional › Finance, accounting & tax | 67.0 | 62.9 | −4.1 |
| AA-PersonalityBench Conviction index | Language & writing › Style & personality | 75.3 | 79.2 | +4.0 |
| Dual-Arm Sim (task completion) | Tool & computer use › Robotics | 7.5 | 3.8 | −3.7 |
| RiemannBench (November 2026 rescoring) | Math | 83.2 | 80.0 | −3.2 |
| Excel Modeling Benchmark v1.1 | Professional › Finance, accounting & tax | 77.4 | 74.3 | −3.1 |
| AutomationBench 1.0.6 (Zapier) | Tool & computer use › Customer service & operations | 44.3 | 41.3 | −3.0 |
| GDP.pdf (November 2026 rescoring) | Professional › General knowledge work | 51.8 | 48.8 | −3.0 |
| DualEntry Accounting AI Benchmark | Professional › Finance, accounting & tax | 81.2 | 84.2 | +3.0 |
| AutomationBench-AA 1.0.6 | Tool & computer use › Customer service & operations | 74.2 | 71.3 | −2.9 |
| Humanity's Sixth Sense | Perception & media › Spatial & visual reasoning | 59.5 | 56.8 | −2.7 |
| GDPval v2 (OpenAI internal) | Professional › General knowledge work | 44.8 | 46.9 | +2.1 |
| Gray Swan IPI Attack Success (k=15) | Safety & alignment | 4.1 | 2.3 | +1.8 |
| Chartography | Perception & media › Documents & charts | 73.4 | 75.1 | +1.7 |
| AA-3D v1.0 Elo | Software › Frontend & design | 1271 | 1259 | −1.7 |
| ARC-AGI-1 Semi-Private | Knowledge & reasoning › Abstract reasoning | 98.5 | 97.5 | −1.0 |
| Ramp SWE-Bench v2 (Bash harness) | Software › Software engineering | 73.0 | 72.4 | −0.6 |
| Blueprint-Bench 2 | Perception & media › Spatial & visual reasoning | 0.6 | 0.6 | +0.4 |
| Agents' Last Exam v1.1 Score | Professional › General knowledge work | 72.8 | 73.0 | +0.3 |
| ARC-AGI-3 (Semi-Private) | Games | 97.3 | 97.4 | +0.1 |
| LMArena Agent Arena Tool Hallucination | Tool & computer use › Tool calling | 1.2 | 1.2 | 0.0 |
Data: Artificial Analysis, Surge AI, Vals AI and 17 othersatlas.village.dev