Nearly half of new Korean AI models build on Chinese foundations

A Rockbridge Network Korea report found 49.3% of the 1,845 models Korean developers published on Hugging Face from January to August were built on Chinese foundations, versus 35.9% on US models. Qwen led at 46.8%, driven by low cost and open weights, raising security and dependency concerns.

• FairStress: New benchmark evaluates fairness in LLM decision-making2026-10-20
• Researchers show edited open models can hide backdoors2026-10-20
• Developer proposes speculative decoding using 2-bit draft plus RAM residual2026-10-22
• Specializing SLM vocabularies can cut inference cost by 77%2026-10-21
• Ornith-1.5-122B released: self-improving model built on Qwen2026-10-20