Supported foundation models in skill-based experience

General considerations

Builders can update the parameters of the model to change the amount of tokens that the output of the model must return in Generative AI skills in Skill studio.

Find information such as the supported languages, latest updates, and training data for the models available in their model card.

Foundation model lifecycle

As new versions of IBM foundation models are introduced, older versions are deprecated. Similarly, as newer and more effective models from other providers become available, less useful models are removed.

For more information on deprecated or withdrawn models, see Deprecated or withdrawn models.

Supported foundation models for different components

The following table shows the list of LLM models that are supported in the product experience of the watsonx Orchestrate by default:

granite-20b-multilingual

IBM Generative AI Large Language foundation models are Enterprise-level language models that are trained with a large volume of data are subjected to intensive pre-processing and careful analysis. The Granite 20 Billion Multilingual (granite-20b-multilingual) model is trained by using over 2.6 trillion tokens and further fine-tuned using a collection of instruction-tuning datasets. The model underwent extended pre-training by using multilingual common crawl data resulting in a model that works with English, German, Spanish, French, and Portuguese.

Usage:

• Question answering
• Summarization
• Retrieval-Augmented Generation
• Classification
• Generation
• Extraction

Model card: ibm/granite-20b-multilingual

granite-13b-instruct-v2

IBM Generative AI Large Language foundation models are Enterprise-level English-language models that are trained with a large volume of data are subjected to intensive pre-processing and careful analysis. The Granite 13 Billion Instruct V2.0 (granite.13b.instruct.v2) model is the instruction-tuned variant that is initialized from the pre-trained Granite Base 13 Billion Base V2.0 (granite.13b.base.v2) model. granite.13b.base.v2 is trained by using over 2.5T tokens. The Granite family of models support all 5 language tasks (Q&A, Generate, Extract, Summarize, and Classify).

Usage:

• Question answering
• Summarization
• Retrieval-Augmented Generation
• Classification
• Generation
• Extraction

Model card: ibm/granite-13b-instruct-v2

granite-3-8b-instruct

Granite-3.0-8B-Instruct is a 8B parameter model fine-tuned from Granite-3.0-8B-Base by using a combination of open source instruction datasets with permissive license and internally collected synthetic datasets. The model is designed to respond to general instructions and can be used to build AI assistants for multiple domains, including business applications.

Usage:

• Question answering
• Summarization
• Classification
• Generation
• Extraction
• Function calling

Model card: ibm/granite-3-8b-instruct

granite-3-2b-instruct

Granite-3.0-2B-Instruct is a lightweight and open source 2B parameter model fine-tuned from Granite-3.0-2B-Base on a combination of open source and proprietary instruction data with a permissively licensed. The model is designed to respond to general instructions and can be used to build AI assistants for multiple domains, including business applications.

Usage:

• Question answering
• Summarization
• Classification
• Generation
• Extraction
• Function calling

Model card: ibm/granite-3-2b-instruct

llama-2-13b-chat

Meta developed and released the Llama 2 family of large language models (LLMs), a collection of pre-trained and fine-tuned generative text models ranging in scale from 7 billion to 70 billion parameters. The fine-tuned LLMs called Llama-2-Chat, are optimized for dialogue use cases. Llama-2-Chat models outperform open source chat models on most benchmarks that were tested, and in the human evaluations for helpfulness and safety, are on par with some closed-source models like ChatGPT and PaLM.

Usage:

• Question answering
• Summarization
• Retrieval-Augmented Generation
• Classification
• Generation
• Code generation and conversion
• Extraction

Model card: meta-llama/llama-2-13b-chat

llama-3-1-70b-instruct

The Meta Llama 3.1 collection of multilingual large language models (LLMs) is a collection of pre-trained and instruction tuned generative models. Instruction tuned text only models are intended for assistant-like chat, whereas pre-trained models can be adapted for various natural language generation tasks. The Llama 3.1 model collection also supports the ability to use the outputs of its models to improve other models including synthetic data generation and distillation.

Usage:

• Question answering
• Summarization
• Retrieval-Augmented Generation
• Classification
• Generation
• Code generation and conversion
• Extraction
• Function calling

Model card: meta-llama/llama-3-1-70b-instruct

llama-3-1-8b-instruct

The Meta Llama 3.1 collection of multilingual large language models (LLMs) is a collection of pretrained and instruction tuned generative models. Instruction tuned text only models are intended for assistant-like chat, whereas pretrained models can be adapted for various natural language generation tasks. The Llama 3.1 model collection also supports the ability to use the outputs of its models to improve other models including synthetic data generation and distillation.

Usage:

• Question answering
• Summarization
• Retrieval-Augmented Generation
• Classification
• Generation
• Code generation and conversion
• Extraction
• Function calling

Model card: meta-llama/llama-3-1-8b-instruct

llama-3-405b-instruct

Llama 3 is intended for commercial and research use in English. Instruction-tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for various natural language generation tasks.

Usage:

• Question answering
• Summarization
• Retrieval-Augmented Generation
• Classification
• Generation
• Code generation and conversion
• Extraction
• Function calling

Model card: meta-llama/llama-3-405b-instruct

llama-3-2-1b-instruct

The Meta Llama 3.2 collection of multilingual large language models (LLMs) is a collection of pretrained and instruction-tuned generative models in 1B and 3B sizes (text in or text out). The Llama 3.2 instruction-tuned text only models are optimized for multilingual dialogue use cases, including agentic retrieval and summarization tasks. They outperform many of the available open sources and closed chat models on common industry benchmarks.

Usage:

• Question answering
• Summarization
• Retrieval-Augmented Generation
• Classification
• Generation
• Code generation and conversion
• Extraction
• Function calling

Model card: meta-llama/llama-3-2-1b-instruct

llama-3-2-3b-instruct

The Meta Llama 3.2 collection of multilingual large language models (LLMs) is a collection of pretrained and instruction-tuned generative models in 1B and 3B sizes (text in/text out). The Llama 3.2 instruction-tuned text only models are optimized for multilingual dialogue use cases, including agentic retrieval and summarization tasks. They outperform many of the available open sources and closed chat models on common industry benchmarks.

Usage:

• Question answering
• Summarization
• Retrieval-Augmented Generation
• Classification
• Generation
• Code generation and conversion
• Extraction
• Function calling

Model card: meta-llama/llama-3-2-3b-instruct

llama-3-2-11b-vision-instruct

The Llama 3.2-Vision collection of multi-modal large language models (LLMs) is a collection of pre-trained and instruction-tuned image reasoning generative models in 11B and 90B sizes (text + images in / text out). The Llama 3.2-Vision instruction-tuned models are optimized for visual recognition, image reasoning, captioning, and answering general questions about an image. The models outperform many of the available open sources and closed multi-modal models on common industry benchmarks.

Usage:

• Question answering
• Summarization
• Retrieval-Augmented Generation
• Classification
• Generation
• Code generation and conversion
• Extraction
• Function calling

Model card: meta-llama/llama-3-2-11b-vision-instruct

llama-3-2-90b-vision-instruct

The Llama 3.2-Vision collection of multi-modal large language models (LLMs) is a collection of pre-trained and instruction-tuned image reasoning generative models in 11B and 90B sizes (text + images in / text out). The Llama 3.2-Vision instruction-tuned models are optimized for visual recognition, image reasoning, captioning, and answering general questions about an image. The models outperform many of the available open sources and closed multi-modal models on common industry benchmarks.

Usage:

• Question answering
• Summarization
• Retrieval-Augmented Generation
• Classification
• Generation
• Code generation and conversion
• Extraction
• Function calling

Model card: meta-llama/llama-3-2-90b-vision-instruct

llama-3-3-70b-instruct

Llama 3.3 is intended for commercial and research use in multiple languages. Instruction tuned text-only models are intended for assistant-like chat, whereas pretrained models can be adapted for various natural language generation tasks.

Usage:

• Question answering
• Summarization
• Retrieval-Augmented Generation
• Classification
• Generation
• Code generation and conversion
• Extraction
• Function calling

Model card: meta-llama/llama-3-3-70B-instruct

mistral-large

Mistral Large 2 is a cutting-edge Large Language Model (LLM) that excels in reasoning, knowledge, and coding domains.

Usage:

• Summarization
• Retrieval-Augmented Generation
• Classification
• Generation
• Code generation and conversion
• Extraction
• Function calling

Model card: mistralai/mistral-large

mixtral-8x7b-instruct-v01

The Mixtral-8x7B Large Language Model (LLM) is a pre-trained generative Sparse Mixture of Experts.

Usage:

• Summarization
• Retrieval-Augmented Generation
• Classification
• Generation
• Code generation and conversion
• Extraction

Model card: mistralai/mixtral-8x7b-instruct-v01

Parent topic:

→ Reference