Over 10,000 engineers powering China’s leading lifestyle services e‑commerce platform. Supporting hundreds of millions of consumers, millions of merchants across 2,000+ industries. This is the public channel for the tech teams behind Meituan, Dianping, Meituan Waimai, Meituan Select, and related services.
This article shares Meituan’s practical approach to controlling the quality of AI‑generated code by using three strategies—unit‑test validation, safety‑net protection for legacy code, and a TDD‑driven workflow—illustrated with real Java examples and detailed test cases.
AMO‑Bench, released by Meituan's LongCat team, is a 50‑question, IMO‑level math reasoning benchmark that combines original, high‑difficulty problems with automated scoring, exposing the current limits of top large language models whose best accuracy hovers around 52 % and offering a more discriminative evaluation tool for future model improvements.
VitaBench, a new open‑source benchmark from Meituan’s LongCat team, evaluates LLM‑driven agents across three realistic life‑service scenarios—food ordering, restaurant dining, and travel planning—using 66 tools and quantifying reasoning, tool, and interaction complexities, exposing a mere 30% success rate on complex cross‑scene tasks.
LongCat-Flash-Omni, the latest open‑source model from Meituan, combines a 560 billion‑parameter architecture, efficient multimodal perception and speech reconstruction modules, and a progressive training strategy to deliver real‑time audio‑video interaction and state‑of‑the‑art performance across text, image, audio, and video tasks.
Meituan’s database team introduced a capacity evaluation system that uses online traffic replay in sandbox environments, accelerated replay to uncover performance bottlenecks, and a capacity‑operation loop to monitor and manage cluster resources, thereby improving database stability, safety of changes, and resource utilization.
This curated list showcases Meituan’s latest large‑model breakthroughs and academic papers up to October 2025, spanning LLM system optimizations, multimodal generation, evaluation benchmarks, quantization techniques, and reinforcement‑learning‑driven improvements, offering researchers valuable insights and resources across the AI landscape.
The paper introduces VSRM, a verifiable step‑reward mechanism that penalizes ineffective reasoning steps and rewards useful ones in large language model inference, dramatically shortening output length while preserving or even improving performance across multiple benchmarks and reinforcement‑learning algorithms.
Meituan’s LongCat team unveiled LongCat-Flash-Thinking, an open‑source large language model that combines deep logical reasoning with tool‑calling capabilities, achieving state‑of‑the‑art performance across logic, mathematics, code, and agentic tasks, and introducing novel training frameworks such as domain‑parallel RL and DORA.
InfiniteTalk introduces a sparse-frame video dubbing paradigm that overcomes traditional mouth‑only editing limits by using keyframe‑guided full‑body generation, employing streaming architecture, soft conditioning, and comprehensive sync to produce seamless, infinite‑length videos for e‑commerce, education, and entertainment.
LongCat-Flash, an open‑source Mixture‑of‑Experts model released by Meituan, leverages model‑system co‑design, PD‑disaggregation, SBO scheduling and large‑scale expert parallelism within the SGLang framework to deliver dramatically lower latency, higher throughput and cost‑effective inference for AI agents, with detailed deployment instructions provided.
