百川智能

Baichuan Intelligence offers an advanced API for large language models tailored for diverse enterprise and development needs. The platform provides access to a range of proprietary models, including Baichuan4-Turbo, Baichuan4-Air, Baichuan4, Baichuan3-Turbo, Baichuan3-Turbo-128k, and Baichuan2-Turbo. These models support up to 192,000 tokens of context window, enabling handling of long documents, extended conversations, and complex reasoning tasks. The API is designed for integration with search enhancement and enterprise knowledge bases, allowing users to combine real-time internet information with internal organizational data for more accurate and context-aware responses. The core functionality revolves around a chat completion endpoint that accepts HTTP POST requests with JSON formatted payloads. Users must include a valid API key in the Authorization header using Bearer authentication. The request body requires specification of the model ID and a list of messages in chronological order. Each message contains a role and content field. Supported roles include system, user, assistant, and tool. The system role is used to set the behavior and context of the assistant, and must be placed as the first message in the conversation history. The content field can be either a plain text string or an array of structured objects enabling multi-modal inputs such as images or URLs, though core text processing remains primary. The platform supports multiple specialized sub-models: a medical large language model tailored for healthcare applications, a general large language model for broad conversational AI, a knowledge base model for retrieval augmented generation, a vectorization model for embedding and similarity search, and a file management model for document processing. This m