Choosing a foundation model
In the agent builder UI, you can select a foundation model for your AI agent.
To do this, locate the AI Model drop-down menu at the top and browse through the available LLM models. Each model offers unique capabilities and characteristics, you can choose the one that suits your specific needs. Refer to the list of Supported foundation models for AI agents for more information.
Each model includes a tag that indicates its status:
Recommended: Model suggested by watsonx Orchestrate
Default: Model currently used by default
Supported foundation models for AI agents
You can use both third-party and IBM foundation models for AI agents in IBM watsonx Orchestrate. The following models are currently supported.
llama-3-2-90b-vision-instruct
The Meta® Llama 3.2 collection of foundation models are provided by Meta. The llama-3-2-90b-vision-instruct models are built for image-in, text-out use cases such as document-level understanding, interpretation of charts and graphs, and captioning of images.
Usage
- Question answering
- Summarization
- Retrieval-Augmented Generation
- Classification
- Generation
- Code generation and conversion
- Extraction
- Function calling
llama-3-405b-instruct
The Llama 3.1 foundation models are pretrained and instruction tuned for text only generation. These models are optimized for multilingual dialogue, with a strong focus on English. Llama 3 supports both commercial and research use.
Usage
- Support chatbot interactions
- Ideal for summarization, content generation, and instruction-following tasks in English