IBM Research reposted this
Today, I’m proud to introduce Granite 3.1 – another step forward in IBM's mission to deliver high-performing, enterprise-ready open source AI models. With every release, we’re getting better and faster – putting out the building blocks businesses need to build reliable, cost-effective AI in a safe, responsible, and scalable way. We’re seeing significant performance improvement with this release – our 8B Instruct model is now topping the Hugging Face OpenLLM V2 leaderboard and competitive with leading lightweight models from Meta, Google, Qwen, and Mistral AI. We’ve expanded the context windows to 128K tokens across the entire Granite family — dense, MoE, and guardrail models — so they can retain and work with long-form content for better performance on general AI tasks. Our Granite Guardian models offer optimal observability with new hallucination detection capability for function-calling, enabling greater reliability and precision in agentic workflows. And above all, we continue to ship under the permissive Apache 2.0 license – reinforcing our belief that AI is nothing if it’s not open, collaborative, and transparent.
Sam Johnston
4d
Apache license or not, the data is the source for AI so Granite is an Open Weight model, not Open Source. Apache may not be a valid license in this context too as it differentiates between "object" and "source" form, and you don't provide the latter i.e. the data. Happy to work with IBM on this in the new year.
Absolutely—openness, collaboration, and transparency are paramount! 🙌 Let’s keep building on that together.
IBM
4d
Raising the bar for lightweight enterprise models ✅
The expanded context windows and hallucination detection capability in the Granite Guardian models are particularly exciting developments.
Pedro Gomez
2d
Para cuando la Singularidad IA-QC, muchas gracias por la generosidad de LA CASA (IBM) en liberar SDK, así se prepara la llegada del NUEVO PARADIGMA ALTA COMPUTACIÓN para la sociedad mundial, DIOS les Bendiga Un admirador desde 1961/ calle serrano de Madrid España
Olcay Dogan
4d
Certainly it looks like more responsible and reliable than its competitors.
Welcome IBM #Granite 3.1 - the reliable and trustworthy Small Language Model for your next #AI solution.
Is the training data going to be publicised? Would be interested to see the jump in quality of the training data between 3 and 3.1!
IBM Senior Vice President and Director of Research.
4dLearn more about today's Granite 3.1 news here: https://www.ibm.com/new/announcements/ibm-granite-3-1-powerful-performance-long-context-and-more