---
title: "Understanding the Market | MEITUAN-W rose over 4% in the afternoon, and the new generation of foundational large model LongCat-2.0-Preview has recently opened for testing"
type: "News"
locale: "en"
url: "https://longbridge.com/en/news/284500449.md"
description: "Meituan-W's stock price rose more than 4% in the afternoon, with an increase of 4.23% as of the time of writing, priced at HKD 83.7, with a transaction volume of HKD 2.045 billion. Meanwhile, Meituan's new generation foundational large model LongCat-2.0-Preview has opened for testing, with a parameter scale exceeding one trillion, achieving world-class levels, supporting complex tasks and enterprise automation applications"
datetime: "2026-04-29T05:51:03.000Z"
locales:
  - [zh-CN](https://longbridge.com/zh-CN/news/284500449.md)
  - [en](https://longbridge.com/en/news/284500449.md)
  - [zh-HK](https://longbridge.com/zh-HK/news/284500449.md)
---

# Understanding the Market | MEITUAN-W rose over 4% in the afternoon, and the new generation of foundational large model LongCat-2.0-Preview has recently opened for testing

According to Zhitong Finance APP, Meituan-W (03690) rose over 4% in the afternoon, with a current increase of 4.23%, priced at HKD 83.7, and a transaction volume of HKD 2.045 billion.

In terms of news, on April 24th, industry sources reported that Meituan-W's new generation foundational large model LongCat-2.0-Preview has opened for testing, with a total parameter scale exceeding one trillion, placing it among the world's top large models. According to insiders, the new generation V4 large model released by DeepSeek on the same day has a total parameter count and active parameter count that are basically consistent with Meituan's LongCat-2.0-Preview.

It is reported that both LongCat-2.0-Preview and DeepSeek V4 support a 1M context window, capable of processing millions of words of input in a single inference, with a processing scale equivalent to the newly released GPT-5.5. Additionally, the new LongCat model has been deeply optimized for Agent application scenarios, effectively adapting to production scenarios such as code generation, complex task planning, and enterprise automation.

Beyond the parameter scale, the greater breakthrough of Meituan's new generation foundational large model lies in its training inference being fully reliant on domestic computing power clusters. According to the aforementioned sources, the number of computing cards used in this training phase by Meituan ranged from 50,000 to 60,000, making it the largest large model training task completed on domestic computing power to date

### Related Stocks

- [03690.HK](https://longbridge.com/en/quote/03690.HK.md)
- [MPNGY.US](https://longbridge.com/en/quote/MPNGY.US.md)
- [DPSK.NA](https://longbridge.com/en/quote/DPSK.NA.md)
- [83690.HK](https://longbridge.com/en/quote/83690.HK.md)
- [HMTD.SG](https://longbridge.com/en/quote/HMTD.SG.md)

## Related News & Research

- [J.P. Morgan Reaffirms Their Hold Rating on Meituan (MPNGF)](https://longbridge.com/en/news/282408172.md)
- [DeepSeek V4 report shows multiple R&D staff departures](https://longbridge.com/en/news/284304672.md)
- [Skymizer Taiwan Inc. Unveils Breakthrough Architecture Enabling Ultra-Large LLM Inference on a Single Card](https://longbridge.com/en/news/283817787.md)
- [Does Robinson’s Nuclear Extension And $103 Billion Grid Plan Reshape The Bull Case For Duke (DUK)?](https://longbridge.com/en/news/284120156.md)
- [DigitalOcean (DOCN) Is Up 6.3% After AI-Driven Q4 Beat And 2026 Guidance Upgrade - Has The Bull Case Changed?](https://longbridge.com/en/news/284160231.md)