Atlas

Compare Models

GPT-5.6 Sol vs Inkling 2 Small

GPT-5.6 Sol leads by 2.7 capability points (≈95% interval +0.6 to +4.8): a clear lead over Inkling 2 Small · P(higher) ≈ 99%

Wins 30 of 50 shared benchmarks (19 lost, 1 tied)

Capability vs. relative cost
GPT-5.6 SolInkling 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.

GPT-5.6 SolvsInkling 2 Smallcapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins30 – 19 · 50 shared, 1 tied

Comparing matching setups only (56 benchmarks excluded) · suites counted once
← GPT-5.6 SolInkling 2 Small →

GPT-5.6 SolvsInkling 2 Smallwins by categoryData Nov 15, 2026Data: Vals AI, Artificial Analysis, Thinking Machines Lab and 12 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 GPT-5.6 Sol.

SRE Bench — Fully SolvedCode30.51.9−28.6
Terminal-Bench 4.0 (Artificial Analysis)Code39.913.1−26.8
EnterpriseBench: CoreCraftProfessional Work46.271.8+25.6
Terminal-Bench 4.0Code37.912.6−25.3
CritPt v2 (agentic, avg@5)Science77.155.7−21.4
Legal Research Bench ImmigrationLegal45.527.3−18.2
Terminal-Bench Science 0.1 (Vals)Science20.02.9−17.1
Terminal-Bench-Science 0.1Science22.45.7−16.7
CrossSearch (pass^3)Search38.023.0−15.0
Legal Research Bench Civil LitigationLegal40.751.9+11.1
ARC-AGI-2General QA92.581.7−10.8
Legal Research Bench HealthLegal80.070.0−10.0
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & Medicine30.020.6−9.4
Legal Research Bench FamilyLegal13.622.7+9.1
Vals Science IndexIndexes41.532.7−8.8
ChartographyMultimodal45.036.7−8.3
Terminal-Bench 2.1Code89.581.3−8.2
MysteryMechanismScience33.325.2−8.1
Design Arena Overall Code CategoriesCode13381381+7.1
Legal Research Bench CriminalLegal43.936.8−7.0
Legal Research Bench Business/CommercialLegal53.546.5−7.0
ScreenSpot-ProMultimodal76.983.8+6.9
ComplexConstraintsChat & Writing50.544.5−6.0
SciCode 1.0.1 (Artificial Analysis)Code57.852.3−5.4
Legal Research Bench Constitutional/Civil RightsLegal34.239.5+5.3
PerceptionBenchMultimodal59.764.7+5.0
BabyVision with PythonMultimodal88.993.6+4.7
EnterpriseOps-Gym-AAAgents42.946.8+3.9
AA-AnalystAgent (pass^5)Professional Work47.551.3+3.7
CyberBench v1.1 OverallCybersecurity76.372.7−3.6
LongBench v2General QA67.170.6+3.5
Humanity's Sixth SenseMultimodal30.033.3+3.3
AA-Briefcase v1.1 EloAgents14791527+2.7
GDP.pdf (Artificial Analysis, All-pass)Professional Work27.825.2−2.6
DAYJOB: Finance (October 2026 rescoring)Finance & Accounting11.814.3+2.5
Tax Agent BenchFinance & Accounting68.065.8−2.2
DeepSearchQA (F1 score)Search93.190.9−2.2
ATLAS-Finance v1.1Finance & Accounting4.06.0+2.0
Humanity's Last Exam DiamondGeneral QA30.428.7−1.7
GPQA DiamondGeneral QA94.693.0−1.6
BioMysteryBench (Vals Terminus 2)Science71.170.0−1.1
Vibe Code Bench v1.1Code80.581.5+1.1
ARC-AGI-1 Semi-PrivateGeneral QA97.596.5−1.0
CharXiv ReasoningMultimodal84.685.6+1.0
ProgramBenchCode1.50.5−1.0
DeepSWE v1.1Code72.772.1−0.6
GDPval-AA v2.1Professional Work16191630+0.5
BrowseCompSearch90.490.2−0.2
Legal Research Bench Weighted PassLegal87.887.9+0.1
Legal Research Bench Administrative/RegulatoryLegal60.660.60.0

GPT-5.6 SolvsInkling 2 Smallnormalised score marginsData Nov 15, 2026Data: Vals AI, Artificial Analysis, Thinking Machines Lab and 12 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…