Ling 3.0 Tiny: benchmark results

Provider: InclusionAI. Released 2026-08-06. Access: Open.

Unified ELO 1497 ± 1, rank #869 of 1761 rated models, from 33 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
AA Omniscience-19.35Score64.3
BenchmarkList ECI120Capability Index (ECI)62.3
Artificial Analysis Intelligence Index15.68Intelligence Index59.9
AA Long Context Reasoning60.33Accuracy (%)55.5
AA GPQA Diamond73.43Accuracy (%)54.1
Tau3 Banking20.82Success Rate (%)54
AA Humanity's Last Exam9.31Accuracy (%)51.9
AA GDPval718.77ELO33.8
AA Omniscience - Software Engineering (SWE) - TypeScript12.22Accuracy (%)32.8
AA-LCR58.67Accuracy (self-reported)27.8
AA Omniscience - Software Engineering (SWE) - PHP14Accuracy (%)27.3
AA Omniscience - Software Engineering (SWE) - Python11.5Accuracy (%)25.1

Interactive version: theaggregate.ai/model?slug=ling-3-0-tiny · How It Works · Data refreshed daily, snapshot 2026-09-05.