---
format: "aidr-story-markdown/v1"
id: "1b25c70569238f706dc5b3f52e244179a98ef69f53a35f363da07dfc67fe0a9b"
canonical_url: "https://aidr.today/1b25c705?lang=en"
title: "LongCat-2.5: Native Multimodality for Agentic Tasks"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-11-02T15:00:26.000Z"
category: "Models"
topics: ["agent"]
source_urls: ["https://longcat.ai/blog/longcat-2.5/","https://news.ycombinator.com/item?id=50337716"]
summary: "LongCat-2.5 is a large model with 1.6万亿 total parameters. The entire training was completed using domestic chips."
---

# LongCat\-2\.5: Native Multimodality for Agentic Tasks

> [Open the canonical story](<https://aidr.today/1b25c705?lang=en>)

**Published:** 2026-11-02T15:00:26.000Z
**Category:** Models
**Topics:** agent

## Summary

LongCat\-2\.5 is a large model with 1\.6万亿 total parameters\. The entire training was completed using domestic chips\.

## Sources

- [Story source](<https://longcat.ai/blog/longcat-2.5/>)
- [Discussion](<https://news.ycombinator.com/item?id=50337716>)