DeepReinforce AI has released Ornith-1.0 on Hugging Face, featuring a diverse range of model architectures and sizes. The collection includes 9B and 31B dense models alongside 35B and 397B mixture-of-experts (MoE) variants. The release claims state-of-the-art performance across various benchmarks, though the validity of these results remains to be seen. Users can access the full collection via the official Hugging Face link provided by the developers. This release expands the available options for large language model inference and fine-tuning.