DeepReinforce AI Releases Ornith-1.0: Open-Source Agentic Coding Models Up to 397B Parameters

Summary

DeepReinforce AI has released Ornith-1.0, a family of open-source models for agentic coding, available in sizes from 9B to 397B parameters. The models are post-trained on top of Gemma 4 and Qwen 3.5, achieving state-of-the-art performance on coding benchmarks like Terminal-Bench 2.1, SWE-Bench, NL2Repo, and OpenClaw. The release features a self-improving training framework using reinforcement learning that jointly optimizes solution generation and the scaffolding that drives those rollouts.

Source

B

bskyDeepReinforce AI Releases Ornith-1.0: Open-Source Agentic Coding Models Up to 397B Parametershuggingface.co