Pangu

Research suggesting Huawei’s Pangu large model shares significant parameter similarity with Alibaba’s Qwen, sparking plagiarism allegations. Huawei’s Pangu team denies this, citing different training hardware and adherence to open-source licenses for shared components. While some code within Pangu bears Qwen’s copyright, it’s attributed to proper utilization of open-source licenses rather than outright plagiarism. The original research report has been removed pending peer review.

Huawei has open-sourced its Pangu 7B dense and Pangu-Pro MoE 72B large language models, along with Ascend-based inference technology. This move supports Huawei’s Ascend ecosystem strategy, aiming to accelerate AI research and application. The Pangu-Pro MoE 72B model shows strong performance, ranking highly on benchmarks for models under 100 billion parameters.