ModelBest CTO Zeng Guoyang: On-device models are the key path for AI deployment

面壁智能CTO曾国洋专访：端侧模型是AI落地关键路径

ModelBest CTO Zeng Guoyang sees on-device models as the key to AI deployment. Their 'model wind tunnel' method predicts full training outcomes from small-scale experiments, and they introduced 'ModelBest's Law': knowledge density doubles every 3.5 months. A 2B-param MiniCPM outperforms same-period 8B competitors. Chips adapted include Qualcomm, MediaTek, Intel, NVIDIA, and AMD. The new BitCPM-CANN series fits roughly 6x more models into the same memory on Huawei Ascend chips. The full-duplex multimodal MiniCPM-o4.5 supports real-time interruption and emotion adjustment. The team also built ForgeTrain, the first production-grade training framework written entirely by AI, and a 'tacit system' that works without speaking via a behavior pattern library.

Why it matters: ModelBest CTO interview with a clear on-device deployment thesis, backed by the 'model wind tunnel' method and 'ModelBest Law' with concrete numbers. The interview format is soft and lacks a sharp edge, so R axis falls short. Score at the featured threshold of 72 without infla...