China model collection

Explore China's leading AI models.

Choose from frontier models for reasoning, code, vision, video, search, and more. Each playground gives you controls tailored to the model.

20 interactive
models

Model directory

Choose a playground

12 organizations represented

Alibaba

6 models

Vision

Qwen2.5-VL-72B-Instruct

Analyze images, documents, charts, and video with a large vision-language model.
Text generation

Qwen3-8B

An efficient text model with optional thinking for reasoning and everyday generation.
Reasoning & agents

Qwen3-235B-A22B

A large mixture-of-experts model for reasoning, multilingual generation, coding, and tools.
Embeddings

Qwen3-Embedding-8B

Create multilingual embeddings for semantic search, retrieval, clustering, and code search.
Reranking

Qwen3-Reranker-8B

Reorder passages, products, code, or documents according to their relevance.
Text to video

Wan2.1-T2V-14B

Generate high-resolution cinematic video from a natural-language description.

Baichuan

1 model

Chat

Baichuan2-Turbo

A commercial chat model for Chinese and English generation, knowledge, and tools.

Baidu

2 models

Reasoning

ERNIE-4.5-21B-A3B-Thinking

An efficient reasoning model for mathematics, science, coding, and long-context analysis.
Document parsing

PaddleOCR-VL

Parse multilingual documents, text, tables, formulas, charts, and page layouts.

ByteDance

1 model

Coding

Seed2.0-Code

Work on software development, code reasoning, debugging, refactoring, and agentic tasks.

DeepSeek

2 models

Reasoning

DeepSeek-R1

A reasoning model for mathematics, software engineering, planning, and analysis.
Text & code

DeepSeek-V3

A general-purpose flagship for conversation, coding, structured output, and tools.

MiniMax

2 models

Long context

MiniMax-Text-01

A long-context model for document analysis, conversation, coding, and tool use.
Role-playing

MiniMax M2-her

Create character-rich role-playing experiences and long-running conversations.

Moonshot AI

1 model

Reasoning & agents

Kimi K2 Instruct

A mixture-of-experts model for coding, research, long-context work, and agents.

OpenGVLab

1 model

Vision

InternVL3-78B

Understand images, charts, documents, spatial relationships, and video.

SenseTime

1 model

Image & text

SenseNova U1

Understand, generate, and edit images with interleaved text and image output.

Tencent

1 model

Hybrid reasoning

Hunyuan-A13B-Instruct

An efficient hybrid-reasoning model for agents, coding, and multilingual chat.

Z.ai

1 model

Reasoning & agents

GLM-4.5

An agent-oriented model for reasoning, coding, web tasks, and function calling.

Zhipu AI

1 model

Text to video

CogVideoX-5B

Turn natural-language scene descriptions into coherent short videos.