The implementation and application of large models have ushered in a critical milestone. The first batch of domestically produced large model products were approved yesterday to be open to the public, with 3 in Shanghai and 5 in Beijing taking the lead in launching them. This means that domestically produced large models have taken an important step from the previous stage of "birth tide" to the new stage of "application tide". What does the big model mean for users to "fly into ordinary people's homes"? What does it mean for domestic large model enterprises? Several experts have stated that the "Battle of Multiple Models" is moving from parameter matching to application matching, which can establish a "flywheel effect" between real user calls and model iterations, which is beneficial for truly serving the Chinese people. The first 8 were approved, and Shanghai occupied 3 seats. Yesterday, the first batch of domestic large model products were approved to open services to users, including Shangtang Technology "Discussion", Baidu "ERNIE Bot", Zhipu AI "Zhipu Qingyan", etc. Baidu first announced the news that "ERNIE Bot" was open to the whole society in the early morning yesterday, which means that users can experience relevant services directly without an internal test invitation code. Compared with the internal test version that was first launched on March 16, the current "ERNIE Bot" has iterated to version 3.5. Even compared to the 3.0 version, the training speed of the new version has increased by 2 times, the inference speed has increased by 17 times, and the cumulative model effect has improved by more than 50%, "said a Baidu insider. Among the first eight large model companies registered through the "Interim Measures for the Management of Generative Artificial Intelligence Services", the reporter noticed that the Shanghai region occupied three seats, namely MiniMax's "ABAB", Shanghai Artificial Intelligence Laboratory's scholar general large model, and Shangtang's "discussion". Among them, "Shangliang" was launched in April this year and is one of the earliest large-scale language models based on hundreds of billions of parameters in China. Before opening to the C-end, it had been applied to various degrees in B-end industries such as finance, healthcare, automotive, real estate, energy, media, and industrial manufacturing. The other two large model companies also have strong technical capabilities. The general model system for scholars is released by the Shanghai Artificial Intelligence Laboratory, where "Scholar Puyu" is also a large language model with billions of parameters. MiniMax, on the other hand, is a "new face", a large model startup established in Shanghai in December 2021. It just received over $250 million in A-round financing in June this year, and its self-developed general model "ABAB" has been approved for release. In addition to Baidu, Tiktok's "Skylark", Baichuan's "Baichuan", Zhipu Huazhang's "Zhipu Qingyan" and the Chinese Academy of Sciences's "Zidong Taichu" are also approved in Beijing. It is reported that three large model companies from provinces and cities such as Guangdong and Anhui will also be launched soon. The "Multi mode Battle" will usher in application layer competition and the "mass launch" of domestic large models, indicating that the "Multi mode Battle" is moving from the previous stage of "birth" to a new stage of "use". Baidu, Tiktok, Shangtang and other enterprises that are moving faster have achieved a "sense of speed" that can be experienced on both mobile and web terminals in just one day. "ERNIE Bot" has even quickly topped the free app ranking list of Apple Store, from which we can see the enthusiasm of users. And "Zhipu Qingyuan" is based on the web and app platforms