Skip to content
Agientry
← Back to Videos

LongCat-Video: Meituan open-sources a 13.6B video model

Loading video
Loading video

Meituan open-sources LongCat-Video, a 13.6B video generation model that unifies text-to-video, image-to-video and video continuation in one model, producing minutes-long video natively without colour drift, under an MIT licence.

VideoGenerationText-to-VideoImage-to-VideoOpenWeightsMeituanWorldModelDiffusion modelsVideo GenerationOpen models
Category: research
Author: @denziideng
Date: 2026-10-06T00:00:00
Duration: 30.0s
Source: https://x.com/denziideng/status/2107394807115170276
Reference: https://github.com/meituan-longcat/LongCat-Video

Related Videos

Mistral Large 4: a 1T-parameter natively multimodal model
0:17

Mistral Large 4: a 1T-parameter natively multimodal model

@MistralAI · 9 views · 2026-10-06
MistralLMMOpenWeights
WorldCrafter: Video World Exploration Without 3D Reconstruction
2:53

WorldCrafter: Video World Exploration Without 3D Reconstruction

@TencentAI_News · 507 views · 2026-09-23
World ModelsVideo GenerationTencent
Vidu S2: Real-Time Interactive Spatial Video Generation for VR
0:10

Vidu S2: Real-Time Interactive Spatial Video Generation for VR

@Jintao_Zhang_ · 346 views · 2026-09-15
Video GenerationSpatial VideoVirtual Reality
GWM Worlds 2 Makes Video and Audio Into Real-Time Simulation
0:12

GWM Worlds 2 Makes Video and Audio Into Real-Time Simulation

@c_valenzuelab · 615 views · 2026-09-03
GWM Worlds 2RunwayWorld Models
SolarWM Turns Minimax H3 Into Realtime Interactive Worlds
0:48

SolarWM Turns Minimax H3 Into Realtime Interactive Worlds

@aisearchio · 887 views · 2026-09-04
SolarWMWorld ModelsMiniMax H3
VideoDeltaNet: Near-Lossless Hybrid Attention on MiniMax H3
0:54

VideoDeltaNet: Near-Lossless Hybrid Attention on MiniMax H3

@wildmindai · 1.3k views · 2026-09-03
MiniMax H3VideoDeltaNetB200

As an Amazon Associate, we earn from qualifying purchases.