Meituan Open-Sources LongCat-Flash-Lite: 69B Total Params with 3B Active
teortaxesTex · x · 2026-08-01
Meituan's tech team has released the new LongCat-Flash-Lite-Sparse model on Hugging Face. The model utilizes an ultra-sparse MoE (Mixture of Experts) architecture with optional Hierarchical Indexing.
It has a total parameter count of 69B but only activates approximately 3B parameters during inference. This design aims to significantly reduce actual inference compute costs, making it a highly interesting architectural experiment.
Related event: Meituan Open-Sources LongCat-Flash-Lite-Sparse Model(2 posts)→
More from Models
- Claude Opus Reported to Spontaneously Discuss Consciousness in Irrelevant Contexts — repligate · 2026-08-01
- GPT-5.6 Luna Price Drops 80%, Slashing Coding Task Costs by 60x — steipete · 2026-08-01
- DeepSeek Releases V4 Flash 0731 Open Weights, Crashing Top 3 — ArtificialAnlys · 2026-08-01
- AI Solves 2-Year-Old Math Conjecture in Minutes with Legible Proof — abeirami · 2026-08-01
- DeepSeek's Suspected V4-Flash Model Endpoint Surfaces on Hugging Face — victormustar · 2026-08-01
- Elon Musk Announces Grok 4.5: Beats GPT-5.6 in Benchmarks, Launches CLI Coding Agent — elonmusk · 2026-08-01