K2-V2 (Low) — benchmark results

Provider: LLM360. Released 2026-01-27. Access: Open.

Unified ELO 1509 ± 15, rank #803 of 1841 rated models, from 53 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
AA Omniscience - Software Engineering (SWE) - Swift36Accuracy (%)47.1
AA LiveCodeBench39.26Pass@1 (%)45.6
AA Omniscience - Humanities & Social Sciences18.6Accuracy (%)44.4
AA Omniscience - Software Engineering (SWE) - Dart18Accuracy (%)40.3
AA IFBench40.95Accuracy (%)40.2
AA Omniscience-48.07Score40.1
AA MMLU-Pro71.29Accuracy (%)38.8
AA-LCR19Score (self-reported)38.3
AA Omniscience - Health16.7Accuracy (%)38.1
AA Omniscience - Software Engineering (SWE) - R8Accuracy (%)37.6
AA Omniscience - Business14.3Accuracy (%)35.9
AA AIME 202535.33Accuracy (%)35.7

Interactive version: theaggregate.ai/model?slug=k2-v2-low · How It Works · Data refreshed daily, snapshot 2026-07-25.