Atlas

Compare Models

Claude Opus 5 vs Inkling 2 Small

Claude Opus 5 leads by 3.9 capability points (≈95% interval +1.7 to +6.2): a clear lead over Inkling 2 Small · P(higher) ≈ 100%

Wins 41 of 49 shared benchmarks (7 lost, 1 tied)

Capability vs. relative cost
Claude Opus 5Inkling 2 SmallClosed 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 Opus 5vsInkling 2 Smallcapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins41 – 7 · 49 shared, 1 tied

Comparing matching setups only (55 benchmarks excluded) · suites counted once
← Claude Opus 5Inkling 2 Small →

Claude Opus 5vsInkling 2 Smallwins by categoryData Nov 15, 2026Data: Vals AI, Artificial Analysis, Surge AI and 16 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Inkling 2 Small minus Claude Opus 5.

Terminal-Bench 4.0 (Artificial Analysis)Code49.013.1−35.9
Terminal-Bench-Science 0.1Science30.05.7−24.3
Terminal-Bench Science 0.1 (Vals)Science27.12.9−24.3
Legal Research Bench FamilyLegal40.922.7−18.2
Legal Research Bench Business/CommercialLegal62.846.5−16.3
APEX-Agents 1.1Professional Work77.060.9−16.1
Humanity's Last Exam DiamondGeneral QA44.828.7−16.1
CritPt v2 (agentic, avg@5)Science71.455.7−15.7
Vals Science IndexIndexes47.932.7−15.2
MysteryMechanismScience37.425.2−12.2
Legal Research Bench Civil LitigationLegal63.051.9−11.1
CrossSearch (pass^3)Search34.023.0−11.0
Tax Agent BenchFinance & Accounting76.765.8−10.9
AA-Omniscience v1.1 IndexIndexes37.616.1−10.8
SWE-Marathon v1.1Code48.838.8−10.0
DAYJOB: Finance (October 2026 rescoring)Finance & Accounting23.814.3−9.5
BioMysteryBench (Vals Terminus 2)Science79.370.0−9.3
ARC-AGI-2General QA90.481.7−8.8
Terminal-Bench 2.1Code89.181.3−7.8
Arena Code WebDev — ReactCode17131659−7.2
ComplexConstraintsChat & Writing37.344.5+7.2
DeepSWE v1.1Code65.072.1+7.1
Legal Research Bench CriminalLegal43.936.8−7.0
AA-Briefcase v1.1 EloAgents16541527−7.0
Vibe Code Bench v1.1Code88.481.5−6.9
Legal Research Bench HealthLegal76.770.0−6.7
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & Medicine13.920.6+6.7
FrontierCode 2.0 - Pass RateCode31.425.5−5.9
FrontierCode 2.0 - ScoreCode28.122.6−5.6
Legal Research Bench Constitutional/Civil RightsLegal44.739.5−5.3
JobBench MainProfessional Work65.770.5+4.8
ATLAS-Finance v1.1Finance & Accounting10.76.0−4.7
Legal Research Bench Administrative/RegulatoryLegal65.260.6−4.5
GDPval-AA v2.1Professional Work17291630−4.4
SciCode 1.0.1 (Artificial Analysis)Code56.452.3−4.1
DeepSearchQA (F1 score)Search94.690.9−3.7
GDP.pdf (Artificial Analysis, All-pass)Professional Work21.625.2+3.6
EnterpriseBench: CoreCraftProfessional Work68.771.8+3.1
Humanity's Sixth SenseMultimodal30.533.3+2.8
Arena Code WebDev — HTMLCode16791660−2.7
Legal Research Bench Weighted PassLegal90.687.9−2.7
AA-AnalystAgent (pass^5)Professional Work53.851.3−2.5
APEX-Accounting 1.1Finance & Accounting64.362.0−2.3
Chartography (with tools)Multimodal84.082.1−1.9
ARC-AGI-1 Semi-PrivateGeneral QA97.596.5−1.0
EnterpriseOps-Gym-AAAgents47.546.8−0.7
MCP Atlas April 2026 (all 1,000 tasks)Agents85.885.4−0.4
CharXiv Reasoning (with tools)Multimodal89.389.1−0.2
Legal Research Bench ImmigrationLegal27.327.30.0

Claude Opus 5vsInkling 2 Smallnormalised score marginsData Nov 15, 2026Data: Vals AI, Artificial Analysis, Surge AI and 16 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…