@fengchris 在 Ornith-1.0系列 Agentic Coding 模型开源发布 9B/31B/35B MoE/397B MoE 四款齐发 中发帖
官博:Ornith-1.0: Self-Scaffolding LLMs for Agentic Coding | DeepReinforce Blog | Jun. 2026
权重:Ornith-1.0 - a deepreinforce-ai Collection
该系列模型基于 Gemma 4 和 Qwen 3.5 构建,各项榜单中同类模型中达到了SOTA水平。
[image]
[image]
[image]
[image]
[image]
强化学习策略:
[image]
PS:也算填补了Qwen-3.7系列暂未开源的空白了吧