IBM announced the launch of its most advanced AI models to date, the Granite 3.0 family, designed to elevate enterprise AI and open-source development, at its annual TechXchange event. This new suite of models—available under the permissive Apache 2.0 license—delivers high performance, flexibility, and safety across a wide range of tasks.
“Granite 3.0: IBM’s third-generation Granite flagship language models can outperform or match similarly sized models from leading providers on many academic and industry benchmarks, showcasing strong performance, transparency, and safety,” IBM said on Monday.
Key Features of Granite 3.0 Models:
General-Purpose Models: The Granite 3.0 8B and 2B language models excel in enterprise tasks such as Retrieval Augmented Generation (RAG), classification, and summarisation, offering up to 23x cost savings compared to larger models.
Safety Focus with Granite Guardian: The new Granite Guardian 3.0 models are tailored to enforce responsible AI practices, checking for risks like bias, toxicity, and hallucinations, ensuring more secure and reliable AI outputs.
“In extensive testing across 19 safety and RAG benchmarks, the Granite Guardian 3.0 8B model has shown higher overall accuracy in harm detection on average than all three generations of Llama Guard models from Meta,” IBM noted.
Mixture of Experts (MoE): MoE models, such as the 1B-A400M and 3B-A800M, are lightweight and optimised for low-latency, CPU-based applications.
Time Series Models: The Granite Time Series models have been upgraded, outperforming larger models from Google, Alibaba, and others, with enhanced flexibility for external variables and rolling forecasts.
According to IBM, the instruct variants of the new Granite 3.0 8B and 2B language models, as well as the Granite Guardian 3.0 8B and 2B models, are available today for commercial use on IBM’s watsonx platform. A selection of the Granite 3.0 models will also be available as NVIDIA NIM microservices and through Google Cloud’s Vertex AI Model Garden integrations with HuggingFace.
IBM’s Granite 3.0 models were trained on over 12 trillion tokens on data taken from 12 different natural languages and 116 different programming languages using two-stage training method. By the end of the year, the 3.0 8B and 2B language models are expected to include support for an extended 128K context window and multi-modal document understanding capabilities.
Assistants to Agents
IBM said it is advancing enterprise AI through a spectrum of technologies—from models and assistants to the tools needed to tune and deploy AI specifically for companies’ unique data and use cases. IBM is also paving the way for future AI agents that can self-direct, reflect, and perform complex tasks in dynamic business environments.
IBM also plans to release new tools to help developers build, customise, and deploy AI more efficiently via watsonx.ai—these include agentic frameworks, integrations with existing environments, and low-code automations for common use cases like RAG and agents.
IBM also announced a major expansion of its AI-powered delivery platform, IBM Consulting Advantage. The multi-model platform includes AI agents, applications, and methods like repeatable frameworks that can empower 160,000 IBM consultants to deliver better and faster client value at a lower cost, according to the official release.
As part of the expansion, Granite 3.0 language models will become the default model in Consulting Advantage.
Another key part of the expansion is the introduction of IBM Consulting Advantage for Cloud Transformation and Management and IBM Consulting Advantage for Business Operations. Each includes domain-specific AI agents, applications, and methods infused with IBM’s best practices so IBM consultants can help accelerate client cloud and AI transformations in tasks, like code modernization and quality engineering, or transform and execute operations across domains, like finance, HR and procurement, IBM said.
If this article saved you time or helped you decide better, consider supporting our work.
Support Telecomtalk
