ZhipuAI Official Website, The Next Generation Cognitive Intelligence Large Model, An AI Company with the Most OpenAI-like Quality and Standards
What is ZhipuAI?
ZhipuAI, derived from the technical achievements of Tsinghua University's Department of Computer Science, is committed to creating the next generation of cognitive intelligence general models. The company has developed the bilingual billion-scale pre-trained model GLM-130B in collaboration and constructed a high-precision general knowledge graph, forming a cognitive engine driven by both data and knowledge. Based on this model, ChatGLM (chatglm.cn) was created. Additionally, ZhipuAI has launched the cognitive large model platform Bigmodel.ai, including products such as CodeGeeX and CogView, providing intelligent API services, connecting billions of users in the physical world, empowering digital humans in the metaverse, and serving as the foundation for embodied robots, endowing machines with the ability to "think" like humans. Based on the leading billion-scale multilingual and multimodal pre-trained model, a new paradigm of AI development is created, which is efficient and generalized "Model as a Service."
Based on the leading billion-scale multilingual and multimodal pre-trained model, creating an efficient and generalized "Model as a Service" AI development new paradigm
ChatGLMSeries Models, Aligned with Human Intentions through Code Pre-training, Supervised Fine-tuning, and Other Techniques, with Capabilities of Q&A, Multi-Round Dialogue, Code Generation, etc.
Bilingual Billion-scale Dialogue Model
Based on the billion-scale base model GLM-130B, infused with code pre-training, and aligned with human intentions through supervised fine-tuning and other techniques, this bilingual large model has functions such as Q&A, multi-round dialogue, and code generation.
AMiner Technology Intelligence Platform
Based on the analysis and mining of three major categories of data: researchers, scientific literature, and academic activities.
Provides academic information retrieval and semantic search, semantic analysis, and achievement evaluation for scientific literature, patents, and science and technology news.
Logical Reasoning
Infused with knowledge and code pre-training, initially possessing reasoning capabilities.
ChatGLM Series ModelsPossessing a broad knowledge base and flexible knowledge association capabilities, the ChatGLM series models have enhanced reasoning capabilities through code pre-training; they can quickly associate a large amount of relevant knowledge and concepts based on input instructions and find the most suitable reasoning chain.
Content Creation
Massive data pre-training, with great potential for creative capabilities.
The ChatGLM large model, based on massive data pre-training, has gained a deep understanding of language, knowledge, and creative techniques, enabling the model to unleash its great potential for creative capabilities and continuously generate rich, extensive, and innovative high-quality original content.
Code Generation
Generate code based on instructions and provide code explanations.
The ChatGLM series models, after code data pre-training, support the rapid generation of code based on natural language prompts and provide code explanations; they also support multiple programming languages and have capabilities such as continuation, translation, annotation, and bugfixing.
Information Extraction
Strong language understanding capabilities, intelligent information extraction.
The ChatGLM series models possess strong language understanding capabilities, capable of deeply understanding the logical relationships between text information and extracting structured information from unstructured text information.
ZhipuAI Code Large Model
CodeGeeX model, with 130 billion parameters, supporting over 20 programming languages, and possessing capabilities such as code generation, continuation, and translation.
