DeepReinforce Releases Four MIT-Licensed Ornith-1.0 Agentic Coding Models

DeepReinforce released Ornith-1.0 on Hugging Face on June 29. The MIT-licensed family includes 9B and 31B dense models and 35B and 397B MoEs, with no regional restrictions. Built on Gemma 4 and Qwen 3.5, the models use reinforcement learning to build their own scaffolding and develop task-specific coding strategies. The 397B MoE scored 77.5 on Terminal-Bench 2.1 and 82.4 on SWE-Bench Verified.

Visit original source