IBM drops open-weight Granite 4.2 family with built-in agentic capabilities under Apache 2.0

IBM has announced the release of its Granite 4.2 language models in sizes of 3B, 8B, and 30B, trained on approximately 15 trillion tokens, with a context window of up to 512,000 tokens. The larger models utilize agentic reinforcement learning to independently learn tool use and code execution. All models are available under the Apache 2.0 license. This launch marks a significant advancement in developing language models capable of performing complex tasks autonomously, enhancing companies' ability to leverage AI in their operations.