IBM (Granite) Open Models: Model List and Licenses

September 27, 2026

About IBM (Granite)

IBM releases Granite, its family of open models for enterprise use, on the ibm-granite organization on Hugging Face. The organization card describes Granite as “a family of open, enterprise-grade AI models that are performant, efficient, and trustworthy".

Granite stands out for its modest sizes and the breadth of its official converted builds, making it one of the easiest publishers to try directly on consumer hardware.

What It Releases

• Granite 4.2 (August 2026): 3B, 8B and 30B language models. According to the model card, this generation builds in reasoning (thinking) natively, significantly improving complex math, coding, multi-step logic and tool-calling tasks. The native context is 128K tokens (extendable to 512K), tested in 12 languages including Japanese.
• Task-specific models: granite-vision-4.1-4b (reads images), granite-guardian-4.1-8b (safety classification), granite-speech-5.0 (speech recognition) and others.
• Research models such as granite-switch and granite-swash are also available.

Licenses

The model card says “all Granite models are released under the Apache 2.0 license, enabling unrestricted commercial and academic use". There are exceptions, though, such as the speech recognition model granite-speech-5.0-470m-turboctc-nc, whose nc marks a non-commercial (CC BY-NC-SA 4.0) license. Read each model’s license text before use. This summary is not legal advice.

Running Them Locally

• Official converted builds are very plentiful. For Granite 4.2, IBM itself publishes GGUF, MLX (for Apple Silicon), FP8, NVFP4 and MXFP4 builds, with many GGUF repositories. They run directly in llama.cpp, Ollama and LM Studio.
• On consumer GPUs, the 3B and 8B are the candidates, and the 30B if you have room. For which quantization fits which GPU, see the memory tables in our articles.

Sources: the ibm-granite organization card; the model card of ibm-granite/granite-4.2-8b; the model list of the ibm-granite organization on Hugging Face (all as of 2026-09-26).

Our Coverage and Data

Local Model Watch has not published an article on IBM (Granite)’s models yet. The model list below is collected by code from Hugging Face every day. Part of our publisher index; model families link to their family pages.

At a Glance

Recent Models on Hugging Face

The publisher lists 241 models in total on Hugging Face. Below are up to 15 of the most recent generative models (text, image, video, audio and so on), the scope of this site, per organization. Classifiers, feature extractors, research components and repositories without a model card are omitted, and the publisher’s own quantized builds and helper variants are folded into the row of the original model. Parameter counts are left blank for repositories that pack weights into integer types, where Hugging Face’s count is not the real parameter count.

From the Hugging Face API. Last changed 2026-09-27 (JST). Collected by code.

Last updated 2026-09-27 (JST). The explanation at the top of this page was written with the help of AI from the primary sources it cites. The tables and lists under “Our Coverage and Data" are assembled by code from our article log and the Hugging Face API.