Qwen1.5-110B, the largest model in the Qwen1.5 series, boasts 11 billion parameters, multilingual support, and an efficient Transformer decoder architecture with Group Query Attention (GQA) for enhanced inference efficiency. It equals Meta-Llama3-70B in baseline performance evaluations and shines in Chat evaluations, including MT-Bench and AlpacaEval 2.0. The release of this model showcases vast potential for performance improvements through scaling model size and indicates that future enhancements will come from expanding data and model scale.