DeepReinforce Releases Ornith-1.0: An Open-Source Coding Model Family That Learns Its Own RL Scaffolds

DeepReinforce has released Ornith-1.0, an open-source coding model family that utilizes a dynamic reinforcement learning approach to generate its own training scaffolds rather than relying on fixed systems. Built on Gemma 4 and Qwen 3.5 architectures, the 397B flagship model achieved an 82.4 score on SWE-Bench Verified. All model weights are available under an MIT license, providing developers with new options for building autonomous coding agents.