LongCat Flash Lite: benchmark results

Meituan LongCat's lite flash tier: 68.5B MoE with ~3B active params and a 256K context (early 2026). Provider: Meituan. Released 2026-01-29. Access: Open.

Unified ELO 1531 ± 1, rank #540 of 1392 rated models, from 37 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
AA TAU-2 Bench79.53Accuracy (%)69.7
ZeroEval MATH-50096.8MATH-500 Score69.4
AA Omniscience - Software Engineering (SWE) - Dart18Accuracy (%)51.5
Artificial Analysis Intelligence Index11.18Intelligence Index50
AA Omniscience - Software Engineering (SWE) - Kotlin16Accuracy (%)47.8
AA IFBench43.06Accuracy (%)46.3
AA Omniscience - Software Engineering (SWE) - Julia8Accuracy (%)46.1
AA Terminal-Bench Hard10.61Accuracy (%)45.1
LLM Stats Score18.91LLM Stats Score (conservative rating)43.9
AA Omniscience - Software Engineering (SWE) - PHP18Accuracy (%)43
BenchmarkList ECI104.41Capability Index (ECI)42.1
LLM Stats (CMMLU)82.48Score (%)40

Interactive version: theaggregate.ai/model?slug=longcat-flash-lite · How It Works · Data refreshed daily, snapshot 2026-09-05.