AIMarkTechPostabout 2 hours ago
DeepReinforce Releases Open-Source Coding Model Ornith-1.0
8 min read
DeepReinforce released Ornith-1.0, an open-source coding model family spanning four sizes from 9B to 397B-MoE. The models learn to write their own reinforcement learning scaffolds instead of using fixed human-designed harnesses. All checkpoints ship under the MIT license on Hugging Face, built on Gemma 4 and Qwen 3.5.
Level
Hype check
Tap to vote and see what everyone thinks.
