Naive AI, founded by Dai Jifeng, released its first open-source model Naive-N0.5-Flash, featuring 309B total parameters and 15.5B activated parameters, native 1M-token context, and a focus on programming and AI research. The model was developed with AI participation and is also trained for AI research work, with the technical report detailing AI's role in attention architecture exploration and training, inference and deployment system optimization.
According to The Information on September 18, 2026, Naive AI, founded by Tsinghua University associate professor Dai Jifeng, has completed three consecutive funding rounds totaling $400 million, with a latest valuation of $1.42 billion. Tencent, IDG Capital, MPCi and HSG have all invested. The team has fewer than 100 people, and its first large language model is nearing release, expected to launch as early as September with open weights.