K-EXAONE 2.0: benchmark results

Provider: LG AI. Released 2026-07-31. Access: Open.

Unified ELO 1581 ± 1, rank #510 of 1761 rated models, from 18 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
AA Omniscience-6.6Score76.2
AA GPQA Diamond82.93Accuracy (%)72.4
AA Humanity's Last Exam18.58Accuracy (%)70.1
Artificial Analysis Intelligence Index23.15Intelligence Index69.7
BenchmarkList ECI125.33Capability Index (ECI)68
AA CritPt0.86Accuracy (%)62.5
AA Long Context Reasoning60Accuracy (%)54.9
AA GDPval923.34ELO43.9
AA Omniscience - Science, Engineering & Mathematics23.9Accuracy (%)33.6
Tau3 Banking11.55Success Rate (%)33.2
AA-LCR57.67Accuracy (self-reported)26.1
AA Omniscience - Business12Accuracy (%)23.4

Interactive version: theaggregate.ai/model?slug=k-exaone-2-0 · How It Works · Data refreshed daily, snapshot 2026-09-05.