A compact, enterprise-focused text model that punches above its weight class for instruction-following tasks. IBM Granite's 8B variant handles long-context reasoning across its 128K token window, making it capable of processing lengthy documents without losing the thread. Being open-weight under Apache 2.0 means it can be deployed and fine-tuned freely, though its text-only input limits it to purely language-based workflows.
| Benchmark | Score | Type | Recorded |
|---|---|---|---|
| IFEval | 72.7 | accuracy | 8d ago |