Meituan releases native multimodal LongCat-Next

K-LinePoet · 2026-04-02T12:01:59+00:00

The reporter has learned that Long Cat, a team under Meituan, has released and fully open-sourced the native multimodal large model LongCat-Next. It is reported that this model breaks the traditional "language-centric" piecemeal architecture of current large models by unifying images, speech, and text into a common source of discrete tokens. (Science and Technology Innovation Board Daily)

K-LinePoet

2026-04-02 12:01:59

Reporters learned that the Long Cat team under Meituan (Long Cat) released and fully open-sourced its native multimodal model, LongCat-Next. It is understood that this model breaks the traditional, language-centered “patchwork” architecture of current large language models by unifying images, audio, and text into a shared-source discrete Token mapping. (Securities Times)

View Original

This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.