Matthias Bastian
IBM has released version 3.1 of its open source Granite LLM, bringing some major improvements under the hood. The new model was trained on a dataset spanning 12 languages and 116 programming languages, processing a total of 12 trillion tokens. The latest version features a redesigned high-density architecture that can handle up to 128,000 tokens at once. According to IBM, these Apache 2.0 license models excel at tasks such as answering questions using external data (RAG), extracting information from unstructured text, and creating document summaries. Developers can now access these models through Hugging Face. IBM first introduced Granite back in May 2024.
Please support our independent, free access reporting. Every contribution helps and secures our future. Support now:
