The Model Index
Up to four models

Qwen3.8-27B vs Inkling-Small

On the index (GPT-4 = 100), Qwen3.8-27B scores 155 and Inkling-Small 153. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.

Next to Qwen3.8-27B
Presets
Qwen3.8-27B
155139–171
WeightsOpen
ReleasedAug 2026
Price in / out$0.060 / $2.2
Context1.02M
Core tests taken3 of 9
Inkling-Small
153137–169
WeightsOpen
ReleasedJul 2026
Price in / out$0.45 / $1.2
Context524K
Core tests taken5 of 9

Qwen3.8-27B and Inkling-Small are too close to call (2 points apart, within the uncertainty), ahead on 2 of 2 shared core tests, at a similar price.

Test by testthird-party results

Qwen3.8-27BInkling-Small
0%25%50%75%100%REASONING & KNOWLEDGEGPQA Diamond (Vals AI)GPQA Diamond (Vals AI)TMMMLU ProMMLU ProTMMATHSProofBench v1.1ProofBench v1.1TMCODINGCode MigrationCode MigrationTMCyberBench v1.1CyberBench v1.1TMIOIIOITMLiveCodeBenchLiveCodeBenchTMProgramBenchProgramBenchTMSWE-benchSWE-benchTMVibe Code Bench v1.1Vibe Code Bench v1.1TMAGENTSSkillsBenchSkillsBenchTMPROFESSIONAL WORKFinance Agent (v2)Finance Agent (v2)TMEMBEMBTMLegalBenchLegalBenchTMLegal Research BenchLegal Research BenchTMMedCodeMedCodeTMMedScribeMedScribeTMMortgageTaxMortgageTaxTMTax Agent BenchTax Agent BenchTMTaxEval v2TaxEval v2TMNOVEL PROBLEMSARC-AGI-2ARC-AGI-2TMARC-AGI-1ARC-AGI-1TM
Every result as a table
TestQwen3.8-27BInkling-Small
Reasoning & knowledge
GPQA Diamond (Vals AI)88.9%83.6%
MMLU Pro84.3%85.6%
Maths
ProofBench v1.116.0%6.0%
Coding
Code Migration14.2%13.7%
CyberBench v1.167.8%64.1%
IOI39.1%9.3%
LiveCodeBench84.0%85.9%
ProgramBench0.0%0.5%
SWE-bench86.0%82.2%
Vibe Code Bench v1.164.8%19.1%
Agents
SkillsBench38.1%33.6%
Professional work
Finance Agent (v2)48.5%41.3%
EMB59.7%31.8%
LegalBench82.4%83.0%
Legal Research Bench36.1%25.5%
MedCode28.7%37.9%
MedScribe83.9%84.1%
MortgageTax64.9%62.3%
Tax Agent Bench30.9%15.4%
TaxEval v270.9%75.5%
Novel problems
ARC-AGI-242.4%40.1%
ARC-AGI-187.5%84.0%

Tests at least two of these models have taken. Bold rows are the core tests.

This page as a chartDownload PNG