K2 Horizon Evaluation Details Show 26% Hallucination Rate, Agent Lead
Artificial Analysis details show K2 Horizon 375B A23B answers only 40% of knowledge questions, refusing rather than guessing, cutting hallucinations to 26%, while leading agent tasks with a GDPval-AA Elo of 1430.
2026-09-03 ~ 2026-09-03 · 4 related posts
- Episode 1: MBZUAI's IFM Open-Sources K2 Horizon Family from 0.9B to 375B(2026-09-03, 6 posts)
- Episode 2: K2 Horizon Evaluation Details Show 26% Hallucination Rate, Agent Lead(2026-09-03, 4 posts)
- K2 Horizon leads agentic peers with GDPval-AA Elo of 1430 vs MiniMax-M3's 1380 — ArtificialAnlys · 2026-09-03
- K2 Horizon declines rather than guesses, cutting hallucination to 26% — ArtificialAnlys · 2026-09-03
- K2 Horizon eval breakdown: declines rather than guesses, 26% hallucination rate — ArtificialAnlys · 2026-09-03
- K2 Horizon 375B A23B Benchmarks Released: Full AI Index Breakdown — ArtificialAnlys · 2026-09-03