LFM2.5-1.2B (Thinking): benchmark results

Provider: Liquid AI. Released 2026-01-20. Access: Open.

Unified ELO 1323 ± 1, rank #1619 of 1761 rated models, from 278 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
EuroEval Spanish NLU - ScaLA ES22.46Linguistic acceptability Score (%)61.8
EuroEval Italian NLU - ScaLA IT22.28Linguistic acceptability Score (%)56.7
EuroEval Portuguese NLU - ScaLA PT15.53Linguistic acceptability Score (%)56.6
AI Chess Leaderboard (Reasoning)721Elo56.5
EuroEval Catalan NLU - ScaLA CA3.31Linguistic acceptability Score (%)46.4
EuroEval Italian Knowledge36.06Knowledge Average Score (%)43.7
AA IFBench41.84Accuracy (%)43.3
EuroEval Spanish NLU - Sentiment Headlines ES35.9Sentiment classification Score (%)43
EuroEval Spanish Knowledge36.8Knowledge Average Score (%)42.7
EuroEval French NLU - ScaLA FR27.82Linguistic acceptability Score (%)42.6
EuroEval Albanian NLU - ScaLA SQ4.46Linguistic acceptability Score (%)42.5
EuroEval English NLU - ScaLA EN38.86Linguistic acceptability Score (%)41.7

Interactive version: theaggregate.ai/model?slug=lfm2-5-1-2b-thinking · How It Works · Data refreshed daily, snapshot 2026-09-05.