IBM Granite is a series of decoder-only AI

foundation models In artificial intelligence (AI), a foundation model (FM), also known as large X model (LxM), is a machine learning or deep learning model trained on vast datasets so that it can be applied across a wide range of use cases.Competition and Markets ...

created by

IBM International Business Machines Corporation (using the trademark IBM), nicknamed Big Blue, is an American Multinational corporation, multinational technology company headquartered in Armonk, New York, and present in over 175 countries. It is ...

. It was announced on September 7, 2023, and an initial paper was published 4 days later. Initially intended for use in the IBM's

cloud-based Cloud computing is "a paradigm for enabling network access to a scalable and elastic pool of shareable physical or virtual resources with self-service provisioning and administration on-demand," according to International Organization for ...

data Data ( , ) are a collection of discrete or continuous values that convey information, describing the quantity, quality, fact, statistics, other basic units of meaning, or simply sequences of symbols that may be further interpreted for ...

and

generative AI Generative artificial intelligence (Generative AI, GenAI, or GAI) is a subfield of artificial intelligence that uses generative models to produce text, images, videos, or other forms of data. These models learn the underlying patterns and str ...

platform

Watsonx Watsonx is IBM's commercial generative AI and scientific data platform based on cloud. It offers a studio, data store, and governance toolkit. It supports multiple large language models (LLMs) along with IBM's own Granite. The platform is de ...

along with other models, IBM opened the source code of some code models. Granite models are trained on datasets curated from

Internet The Internet (or internet) is the Global network, global system of interconnected computer networks that uses the Internet protocol suite (TCP/IP) to communicate between networks and devices. It is a internetworking, network of networks ...

, academic publishings,

code In communications and information processing, code is a system of rules to convert information—such as a letter, word, sound, image, or gesture—into another form, sometimes shortened or secret, for communication through a communicati ...

datasets,

legal Law is a set of rules that are created and are law enforcement, enforceable by social or governmental institutions to regulate behavior, with its precise definition a matter of longstanding debate. It has been variously described as a Socia ...

and finance documents.

Foundation models

A foundation model is an AI model trained on broad data at scale such that it can be adapted to a wide range of downstream tasks. Granite's first foundation models were Granite.13b.instruct and Granite.13b.chat. The "13b" in their name comes from 13 billion, the amount of

parameters A parameter (), generally, is any characteristic that can help in defining or classifying a particular system (meaning an event, project, object, situation, etc.). That is, a parameter is an element of a system that is useful, or critical, when ...

they have as models, lesser than most of the larger models of the time. Later models vary from 3 to 34 billion parameters. On May 6, 2024, IBM released the

source code In computing, source code, or simply code or source, is a plain text computer program written in a programming language. A programmer writes the human readable source code to control the behavior of a computer. Since a computer, at base, only ...

of four variations of Granite Code Models under

Apache 2 APACHE II ("Acute Physiology and Chronic Health Evaluation II") is a severity-of-disease classification system, one of several ICU scoring systems. It is applied within 24 hours of admission of a patient to an intensive care unit (ICU): an in ...

, an open source

permissive license A permissive software license, sometimes also called BSD-like or BSD-style license, is a free-software license which instead of copyleft protections, carries only minimal restrictions on how the software can be used, modified, and redistributed, ...

that allows completely free use, modification and sharing of the software, and put them on

Hugging Face Hugging Face, Inc. is a French-American company based in List of tech companies in the New York metropolitan area, New York City that develops computation tools for building applications using machine learning. It is most notable for its Transf ...

for public use. According to IBM's own report, Granite 8b outperforms

Llama 3 Llama (Large Language Model Meta AI, formerly stylized as LLaMA) is a family of large language models (LLMs) released by Meta AI starting in February 2023. The latest version is Llama 4, released in April 2025. Llama models come in different s ...

on several coding related tasks within similar range of parameters.

References

External links

GitHub page

IBM Granite Playground
{{Generative AI IBM products IBM software Large language models Generative artificial intelligence Artificial neural networks 2023 software Free software Open-source artificial intelligence

Foundation models

See also

References

External links