A lightweight model that punches at its weight class — which is quite small. With only 0.6B parameters, it handles simple text tasks and basic comprehension but struggles with complex reasoning or nuanced instructions. Its 32K context window is surprisingly generous for its size, making it useful for processing longer inputs even if deep understanding remains limited.
| Benchmark | Score | Type | Recorded |
|---|---|---|---|
| IFEval | 17.1 | accuracy | today |