Granite Team: transparent enterprise-oriented language-model training resource

What changed

Released compact dense and mixture-of-experts models with documented two-stage data-mixture selection, training and post-training procedures, supporting inspectable enterprise-oriented customization.

Scientific scope

Released compact dense and mixture-of-experts models with documented two-stage data-mixture selection, training and post-training procedures, supporting inspectable enterprise-oriented customization. An integrated training/resource contribution, not invention of inherited PPO or mixture-of-experts methods. Data governance procedures and reported benchmark performance do not guarantee legal compliance, safety or superiority. Later 128K/multimodal plans are excluded.

Evidence

The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. An integrated training/resource contribution, not invention of inherited PPO or mixture-of-experts methods. Data governance procedures and reported benchmark performance do not guarantee legal compliance, safety or superiority. Later 128K/multimodal plans are excluded.

Contributors

• Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.

• Granite 3.0 Language Models
• Granite Team: transparent enterprise-oriented language-model training resource — institutional/author source

Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.

Full scientific record · Browse contribution grades · Request a correction