A text-to-text generalist that frames every NLP task — translation, summarization, classification, question answering — as a sequence-to-sequence problem. With 11 billion parameters, it's the largest member of the T5 family, offering strong transfer learning across diverse language tasks. It requires significant compute to run and fine-tune, which is a practical constraint for resource-limited environments.