LFM 40B — benchmark results

Provider: Liquid AI. Released 2024-10-23. Access: Open.

Unified ELO 1327 ± 50, rank #1479 of 1776 rated models, from 8 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
AA Humanity's Last Exam4.88Accuracy (%)30.7
Wolfram LLM Benchmarking Project22.7Correct Functionality (%)21.1
AA MATH-50048Accuracy (%)12.9
AA GPQA Diamond32.73Accuracy (%)10.6
AA MMLU-Pro42.53Accuracy (%)9.3
Artificial Analysis Intelligence Index3.37Intelligence Index8.9
AA LiveCodeBench9.63Pass@1 (%)7.2
AA SciCode7.06Accuracy (%)7.1

Interactive version: theaggregate.ai/model?slug=lfm-40b · How the rankings work · Data refreshed daily, snapshot 2026-07-22.