"IBM has launched Granite 4.0, a new family of open-source language models designed to slash infrastructure costs that have become a major barrier to enterprise AI adoption. Released under the Apache 2.0 licensing, Granite 4.0 represents IBM's bet on a fundamentally different architectural approach to enterprise AI deployment. The models are built on what the company described as a "hybrid" architecture - combining emerging Mamba state space models with traditional transformer layers."
"The release included base and instruction-tuned variants across three primary models: Granite-4.0-H-Small (32 billion total parameters, 9 billion active), Granite-4.0-H-Tiny (7 billion total, 1 billion active), and Granite-4.0-H-Micro (3 billion dense). IBM said the Tiny and Micro models are "designed for low latency, edge, and local applications." "Relative to conventional LLMs, our hybrid Granite 4.0 models require significantly less RAM to run, especially for tasks involving long context lengths (like ingesting a large codebase or extensive documentation)""
Granite 4.0 is an open-source family of language models released under Apache 2.0 that targets enterprise infrastructure cost reduction. The models use a hybrid architecture combining Mamba state-space layers with traditional transformers, enabling linear computational scaling with sequence length instead of transformers' quadratic scaling. Granite 4.0 comes in base and instruction-tuned variants across three sizes: H-Small (32B total, 9B active), H-Tiny (7B total, 1B active), and H-Micro (3B dense), with Tiny and Micro optimized for low-latency, edge, and local deployment. The hybrid approach reduces RAM requirements for long-context and multi-session tasks, accelerates inference, and carries ISO 42001 certification for trust.
#mamba-state-space #hybrid-mamba-transformer #long-context-efficiency #edge-deployment #apache-20-open-source
Read at InfoWorld
Unable to calculate read time
Collection
[
|
...
]