Atlas

Compare Models

Gemini 3.8 Flash vs Inkling 2 Small

Gemini 3.8 Flash leads by 0.5 capability points (≈95% interval −1.6 to +2.7): statistically tied · P(higher) ≈ 69%

Inkling 2 Small wins 56 of 107 shared benchmarks (43 lost, 8 tied)

Capability vs. relative cost
Gemini 3.8 FlashInkling 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.

Gemini 3.8 FlashvsInkling 2 Smallcapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins43 – 56 · 107 shared, 8 tied

Comparing matching setups only (53 benchmarks excluded) · suites counted once
← Gemini 3.8 FlashInkling 2 Small →

Gemini 3.8 FlashvsInkling 2 Smallwins by categoryData Nov 15, 2026Data: Vals AI, Artificial Analysis, Surge AI and 10 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 Gemini 3.8 Flash.

HLAB-Vals-1.1 - Antitrust CompetitionLegal0.040.0+40.0
BrokenArXiv 09/2026 (Native Harnesses)Math26.763.2+36.5
BrokenArXiv 08/2026 (Native Harnesses)Math27.463.1+35.7
HLAB-Vals-1.1 - Arbitration International Dispute ResolutionLegal0.030.0+30.0
HLAB-Vals-1.1 - Energy Natural ResourcesLegal20.050.0+30.0
HLAB-Vals-1.1 - Funds Asset ManagementLegal0.030.0+30.0
HLAB-Vals-1.1 - ImmigrationLegal0.030.0+30.0
HLAB-Vals-1.1 - Litigation Dispute ResolutionLegal10.040.0+30.0
CyberBench v1.1 OverallCybersecurity43.872.7+29.0
Legal Research Bench Civil LitigationLegal29.651.9+22.2
HLAB-Vals-1.1 - Bankruptcy RestructuringLegal0.020.0+20.0
HLAB-Vals-1.1 - Real EstateLegal10.030.0+20.0
HLAB-Vals-1.1 - Structured Finance SecuritizationLegal0.020.0+20.0
ArXivMath 09/2026 (Native Harnesses)Math50.031.1−18.9
AA-Briefcase v1.1 EloAgents12041527+17.7
Legal Research Bench Administrative/RegulatoryLegal47.060.6+13.6
Legal Research Bench HealthLegal56.770.0+13.3
EnterpriseBench: CoreCraftProfessional Work58.571.8+13.3
APEX-Agents 1.1Professional Work72.960.9−12.0
Finance Agent v2.1 - Market AnalysisFinance & Accounting70.559.0−11.5
Arena Code WebDev — ReactCode15751659+11.3
MysteryMechanismScience36.525.2−11.3
HLAB-Vals-1.1 - Banking FinanceLegal0.010.0+10.0
HLAB-Vals-1.1 - Capital MarketsLegal30.020.0−10.0
HLAB-Vals-1.1 - Corporate GovernanceLegal20.030.0+10.0
HLAB-Vals-1.1 - Corporate M&ALegal30.020.0−10.0
HLAB-Vals-1.1 - Emerging Companies Venture CapitalLegal30.020.0−10.0
HLAB-Vals-1.1 - Intellectual PropertyLegal30.020.0−10.0
HLAB-Vals-1.1 - TaxLegal10.020.0+10.0
HLAB-Vals-1.1 - Trusts Estates Private ClientLegal30.020.0−10.0
DAYJOB: Finance (October 2026 rescoring)Finance & Accounting4.514.3+9.8
CritPt v2 (agentic, avg@5)Science65.155.7−9.4
HANDBOOK.md Agents v1.1Professional Work26.517.8−8.7
Humanity's Sixth SenseMultimodal41.633.3−8.3
GDPval-AA v2.1Professional Work14421630+8.3
Finance Agent v2.1 - PrecedentsFinance & Accounting38.947.1+8.2
VIALS-AA (pass@3)Multimodal64.072.0+8.1
BioMysteryBench (Vals Terminus 2)Science62.270.0+7.8
AA-3D v1.0 EloCode9741027+7.7
ARC-AGI-2General QA89.281.7−7.5
Arena Code WebDev — HTMLCode16091660+7.1
Terminal-Bench 4.0Code19.712.6−7.1
Finance Agent v2.1 - General Qualitative AnalysisFinance & Accounting71.078.0+7.0
FrontierCode 2.0 (Artificial Analysis)Code18.725.4+6.7
Terminal-Bench 4.0 (Artificial Analysis)Code19.713.1−6.6
Terminal-Bench 2.1Code87.681.3−6.3
Excel Modeling Benchmark v1.1 LBOFinance & Accounting62.255.9−6.3
Excel Modeling Benchmark v1.1 M&AFinance & Accounting79.973.7−6.2
CrossSearch (pass^3)Search29.023.0−6.0
VIALS-AA (average of 3)Multimodal52.057.8+5.8
Finance Agent v2.1 - ComparablesFinance & Accounting59.265.0+5.7
Terminal-Bench Science 0.1 (Vals)Science8.62.9−5.7
Finance Agent v2.1 - AdjustmentsFinance & Accounting61.967.4+5.5
Excel Modeling Benchmark v1.1 TemplateFinance & Accounting76.971.6−5.3
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & Medicine15.620.6+5.0
AA-AnalystAgent (pass^5)Professional Work46.351.3+5.0
Terminal-Bench-Science 0.1Science10.55.7−4.8
Legal Research Bench Business/CommercialLegal41.946.5+4.7
SWE Atlas v1.1 - OverallCode57.361.9+4.6
SWE-bench MultilingualCode85.881.2−4.6
Legal Research Bench FamilyLegal18.222.7+4.5
Excel Modeling Benchmark v1.1 DataroomFinance & Accounting87.883.4−4.4
SciCode 1.0.1 (Artificial Analysis)Code56.652.3−4.3
Finance Agent v2.1 - Disclosure AnalysisFinance & Accounting76.680.8+4.2
ChartographyMultimodal40.936.7−4.2
ComplexConstraintsChat & Writing48.444.5−3.9
Excel Modeling Benchmark v1.1 DCFFinance & Accounting66.162.2−3.9
AA-Omniscience v1.1 IndexIndexes23.716.1−3.8
SAGE v1.1 - Linear AlgebraMath84.080.5−3.5
ArXivMath 08/2026 (Native Harnesses)Math40.436.8−3.5
Legal Research Bench CriminalLegal40.436.8−3.5
sudo L7 (October 2026 rescoring)Code10.614.1+3.5
APEX-Accounting 1.1Finance & Accounting58.862.0+3.2
Vals Science IndexIndexes35.832.7−3.1
Excel Modeling Benchmark v1.1 CompsFinance & Accounting76.073.1−2.9
Vibe Code Bench v1.1Code78.781.5+2.9
SAGE v1.1 - Partial Differential EquationsMath91.194.0+2.9
Finance Agent v2.1 - Financial ModelingFinance & Accounting45.142.4−2.7
Legal Research Bench Constitutional/Civil RightsLegal42.139.5−2.6
GDP.pdf (Artificial Analysis, All-pass)Professional Work22.825.2+2.4
ATLAS-Finance v1.1Finance & Accounting3.76.0+2.3
SWE-bench Multimodal (Vals v1, 480 Tasks)Code40.843.1+2.3
Finance Agent v2.1 - All PassFinance & Accounting54.556.6+2.1
ARC-AGI-1 Semi-PrivateGeneral QA98.596.5−2.0
VIALS-AA (pass^3)Multimodal41.042.9+1.9
SAGE v1.1 - Computer ScienceMath89.691.2+1.6
SWE-bench VerifiedCode80.682.2+1.6
Excel Modeling Benchmark v1.1 ScratchFinance & Accounting68.567.1−1.5
Code Migration v1.1 (Vals Index v2.2 subset)Code38.239.5+1.3
HLAB-Vals-1.1 - Criteria Pass RateLegal92.693.7+1.1
DeepSWE v1.1Code71.072.1+1.1
Tax Agent BenchFinance & Accounting66.865.8−1.0
Finance Agent v2.1 - General Quantitative AnalysisFinance & Accounting80.279.2−1.0
Excel Modeling Benchmark v1.1 OperatingFinance & Accounting69.270.0+0.8
Excel Modeling Benchmark v1.1 MiscFinance & Accounting68.968.3−0.6
Finance Agent v2.1 - Earnings AnalysisFinance & Accounting88.488.9+0.5
ProgramBenchCode1.00.5−0.5
Legal Research Bench Weighted PassLegal87.587.9+0.4
Humanity's Last Exam DiamondGeneral QA29.128.7−0.4
HLAB-Vals-1.1 - Data Privacy CybersecurityLegal20.020.00.0
HLAB-Vals-1.1 - Employment LaborLegal0.00.00.0
HLAB-Vals-1.1 - Environmental ESGLegal0.00.00.0
HLAB-Vals-1.1 - Healthcare Life SciencesLegal30.030.00.0
HLAB-Vals-1.1 - InsuranceLegal0.00.00.0
HLAB-Vals-1.1 - International Trade SanctionsLegal20.020.00.0
HLAB-Vals-1.1 - White Collar Defense InvestigationsLegal0.00.00.0
Legal Research Bench ImmigrationLegal27.327.30.0

Gemini 3.8 FlashvsInkling 2 Smallnormalised score marginsData Nov 15, 2026Data: Vals AI, Artificial Analysis, Surge AI and 10 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…