Ornith-1.0: Self-Scaffolding LLMs for Agentic Coding
DeepReinforce has released Ornith-1.0, an open-weights (MIT licensed) family of self-scaffolding LLMs for agentic coding, built on Gemma 4 and Qwen 3.5 with variants up to 397B parameters. Initial testing with the 35B GGUF model shows strong performance on coding agent tasks, running proficiently through many tool calls.
Background
- **Ornith-1.0** is a new family of open-source coding AI models released by a startup called **DeepReinforce**. The models come in sizes from 9 billion to 397 billion parameters, and are "self-scaffolding" — meaning they are designed to autonomously use tools, call functions, and manage multi-step coding tasks without constant human prompting. This makes them useful for "agentic" coding (letting an AI act as an autonomous coding assistant).