Skip to main content

Overview

Here you will find a quick view of available ready-to-use integration capabilities and all our supported integrations and models for every LLM provider supported on LangChain4j so far.

We are making a great effort to have most of the functions enabled according to the progress and updates of the LLM providers and java features.

Capabilities​

  1. Native image: You can use this LLM integration for AOT compilation using GraalVM CE or GraalVM Oracle for native image generation.
  2. Sync Completion: Supports the implementation of text-completion and chat-completion models in a synchronous way. This is the most common usage.
  3. Streaming Completion: Supports streaming the model response back for text-completion or chat-completion models, handling each event in StreamingResponseHandler<AiMessage> class. View examples here
  4. Embeddings: Supports the implementation of text-embedding models. Embeddings make it easy to add custom data without fine-tuning. Generally used with RAG (Retrieval-Augmented Generation) and Embedding Stores.
  5. Image Generation: Supports the implementation of text-to-image models to create realistic and coherent images from scratch. View examples here
  6. Scoring: Understands the implementation of scoring models to improve created models by re-organizing their results based on certain parameters.
  7. Function Calling: Supports the implementation of function-calling models to call a function as a Tool. View examples here
note

of course some LLM providers offer large multimodal model (accepting text or image inputs) and it would cover more than one capability.

Supported LLM Integrations​

ProviderNative ImageSync CompletionStreaming CompletionEmbeddingImage GenerationScoringFunction Calling
OpenAI✅✅✅✅✅
Azure OpenAI✅✅✅✅✅
Hugging Face✅✅
Amazon Bedrock✅✅✅
Google Vertex AI Gemini✅✅✅✅
Google Vertex AI✅✅✅✅
Mistral AI✅✅✅✅
DashScope✅✅✅
LocalAI✅✅✅✅
Ollama✅✅✅
Cohere✅
Qianfan✅✅✅✅
ChatGLM✅
Nomic✅
Anthropic✅✅
Zhipu AI✅✅✅✅