Asian AI model index

Models built in Asia, explained clearly.

Browse 45 AI models from China, Japan, South Korea, India and Southeast Asia. Each profile includes task coverage, access notes, editorial scores, a review date and official sources.

All models · Page 2 of 4

Sorted by AIOAI’s overall discovery score. Use a task ranking when you need a more specific answer.

13

HappyHorse

Alibaba ATH's first-party image-to-video model, exposed through Model Studio for motion-rich video generation from a starting frame.

Alibaba ATH · China89Estimate
14

Step

A multimodal model family covering reasoning, coding, image, video and speech generation.

StepFun · China89Estimate
15

Hunyuan

Tencent's model family for text, code, images, video and enterprise applications.

Tencent · China88Estimate
16

Hailuo AI

An accessible AI video generator for short clips, characters and visual concepts.

MiniMax · China88Estimate
17

SkyReels

Kunlun Tech's first-party video model family and creator platform, including openly released human-centric video models.

Kunlun Tech · China88Estimate
18

MuseSteamer

Baidu's first-party video-generation model available through Qianfan for text-to-video and image-to-video production workflows.

Baidu · China88Estimate
19

EXAONE

LG's bilingual and multimodal model family for professional and research use.

LG AI Research · South Korea88Estimate
20

Kanana

Kakao's Korean-first language-model family with open releases for chat, reasoning and developer use.

Kakao · South Korea88Estimate
21

AI Studios

DeepBrain AI's first-party avatar-video platform for scripted presenters, multilingual dubbing, training and business communication.

DeepBrain AI · South Korea88Estimate
22

ILMU

A Malaysian multimodal model family built around local languages, cultural knowledge and regional deployment.

YTL AI Labs · Southeast Asia88Estimate
23

CogVideoX

Zhipu AI's open video-generation model series for text-to-video and image-to-video research, deployment and API use.

Z.ai · Zhipu AI · China87Estimate
24

HyperCLOVA X

A Korean-first foundation model designed for language, reasoning and enterprise use.

NAVER Cloud · South Korea87Estimate
Coverage last reviewed 2026-08-21. This is a focused editorial dataset, not an exhaustive list of every model in Asia. See how models are selected and scored.