Atlas

Compare Models

Claude Opus 5.5 vs GPT-6.2 Sol

Claude Opus 5.5 leads by 0.7 capability points (≈95% interval −2.1 to +3.4): statistically tied · P(higher) ≈ 68%

Wins 35 of 55 shared benchmarks (18 lost, 2 tied)

Capability vs. relative cost
Claude Opus 5.5GPT-6.2 SolClosed weightsOpen weights

Benchmark wins35 – 18 · 55 shared, 2 tied

Comparing matching setups only (51 benchmarks excluded) · suites counted once
← Claude Opus 5.5GPT-6.2 Sol →
Data: Artificial Analysis, Surge AI, Arena and 15 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 Opus 5.5.

AA-PersonalityBench Informativeness indexLanguage & writing › Style & personality90.537.7−53.3
TxBench Preclinical Pharmacology V1 Clean RerunsScience › Life sciences26.365.3+39.0
EnterpriseBench: CoreCraftTool & computer use › Customer service & operations74.436.4−38.0
AA-PersonalityBench Expressiveness indexLanguage & writing › Style & personality90.555.3−35.5
AA-PersonalityBench Imagination indexLanguage & writing › Style & personality94.360.1−34.6
AA-PersonalityBench Transparency indexLanguage & writing › Style & personality90.556.9−33.9
AA-PersonalityBench Organization indexLanguage & writing › Style & personality79.145.7−33.7
DAYJOB: Healthcare (October 2026 rescoring)Health & medicine48.317.2−31.1
AA-PersonalityBench Warmth indexLanguage & writing › Style & personality84.860.1−24.9
DeepsecBench — Revalidated V1 (233 issues)Software › Cybersecurity26.647.3+20.8
SimpleBenchKnowledge & reasoning › General knowledge88.469.9−18.5
HANDBOOK.md Agents v1.1Tool & computer use › Customer service & operations52.334.5−17.8
SWE Atlas v1.1 - OverallSoftware › Software engineering77.560.5−17.0
LMArena Agent Arena Net Improvement (Steering Burden)Tool & computer use › General agents18.611.3−14.3
Humanity's Sixth SensePerception & media › Spatial & visual reasoning44.656.8+12.2
AA-Briefcase v1.1 EloProfessional › General knowledge work18001612−10.3
sudo L7 (October 2026 rescoring)Software › Software engineering46.136.1−10.0
TaxCalcBench TY25Professional › Finance, accounting & tax42.032.0−10.0
AA-PersonalityBench Conviction indexLanguage & writing › Style & personality69.579.2+9.8
DAYJOB: Finance (October 2026 rescoring)Professional › Finance, accounting & tax42.532.8−9.7
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & medicine27.837.2+9.4
Humanity's Last Exam DiamondKnowledge & reasoning › General knowledge54.664.0+9.4
DAYJOB: Law (November 2026 release)Professional › Legal23.732.7+9.0
GDP.xlsxProfessional › General knowledge work30.322.3−8.0
AA-3D v1.0 EloSoftware › Frontend & design12021259+7.7
LMArena Agent Arena Praise vs ComplaintTool & computer use › General agents38.623.3−7.7
ARC-AGI-2Knowledge & reasoning › Abstract reasoning93.385.8−7.5
DAYJOB: AccountingProfessional › Finance, accounting & tax34.427.6−6.8
Blueprint-Bench 2Perception & media › Spatial & visual reasoning0.50.6+5.8
ComplexConstraints (November 2026 rescoring)Language & writing › Instruction following64.469.5+5.1
LMArena Agent Arena Bash ReliabilityTool & computer use › General agents14.04.3−4.9
Terminal-Bench-Science 0.1Science › Multidisciplinary61.966.7+4.8
AA-PersonalityBench Slop indexLanguage & writing › Style & personality20.015.4−4.7
DAYJOB: TaxProfessional › Finance, accounting & tax9.24.8−4.4
Professional Reasoning Benchmark - LegalProfessional › Legal47.643.3−4.2
LMArena Agent Arena Confirmed SuccessTool & computer use › General agents19.210.9−4.2
Dual-Arm Sim (task completion)Tool & computer use › Robotics0.03.8+3.8
Terminal-Bench 4.0 (Artificial Analysis)Software › Software engineering59.656.1−3.5
GDPval v2 (OpenAI internal)Professional › General knowledge work43.546.9+3.4
CritPt v2 (agentic, avg@5)Science › Physical sciences81.778.6−3.1
LiveBench 2.0 (2026-11-06 release)Composite indexes83.480.3−3.1
HLAB-Vals-1.1Professional › Legal9.212.1+2.9
AA-AnalystAgent (pass^5)Professional › General knowledge work56.353.8−2.5
LMArena Agent Arena Steering BurdenTool & computer use › General agents18.013.4−2.3
FrontierCode 2.0 (Artificial Analysis)Software › Software engineering46.444.4−2.1
DualEntry Accounting AI BenchmarkProfessional › Finance, accounting & tax82.284.2+2.0
Chartography (November 2026 rescoring)Perception & media › Documents & charts89.687.7−1.9
Excel Modeling Benchmark v1.1Professional › Finance, accounting & tax75.874.3−1.5
Gray Swan IPI Attack Success (k=15)Safety & alignment1.02.3−1.3
Ramp SWE-Bench v2 (Bash harness)Software › Software engineering71.272.4+1.2
GDP.pdf (Artificial Analysis, All-pass)Professional › General knowledge work28.830.0+1.2
ARC-AGI-1 Semi-PrivateKnowledge & reasoning › Abstract reasoning98.597.5−1.0
LMArena Agent Arena Tool HallucinationTool & computer use › Tool calling1.01.2+0.1
GDP.pdf (November 2026 rescoring)Professional › General knowledge work48.848.80.0
RiemannBench (November 2026 rescoring)Math80.080.00.0
Data: Artificial Analysis, Surge AI, Arena and 15 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…