A large open-weight text model from ornith-ai, Ornith 1.5 122B FP8 runs in a quantized FP8 format that trades a small amount of numerical precision for reduced memory footprint and faster inference. At 122 billion parameters it sits in the heavyweight category, suggesting capacity for complex reasoning and nuanced language tasks, though specific capability details beyond text-in, text-out are not documented. The MIT license makes it freely usable and modifiable.