[Interpreted] Thanks for the question. Regarding the existing customer, yes, indeed, for the existing customer or the user, we do have a very high retention rate because many of them are our long-term users, and if we're going to convert them, that means they just continue to renew our service. So mechanistic in the conversion rate of our existing users would be very high. Well, for the newly acquired users, as we mentioned just now, in order to engage those users we provide some discounts and some promotional activities even if we are already scaling back the discount, but actually, the conversion rate of the new users will go in line with our promotional activities and sometimes it's subject to the change of the promotional activity time line.
The second part of our question is regarding how AI can empower user retention. I have to say that the large model indeed helped to ground the retention. But let's be clear first. For the music platform, the recommendation system was based on the newer network model, and it is different from the large language model we're talking about today. But generously speaking, I do believe as we're having those great models and by introducing the large set of parameters into our existing platform, we will be able to continue to improve the recommendation capacity and search capacity. As you can see from our actual operationals, that can help us to continue to grow and optimize the US retention. So indeed, by leveraging those caring technologies, our overall retention rates improved.
Actually, when you talk about large models and generated technology. In Q1 of this year, we also launched the open-source of self-developed minicam video model drives, which has already seen very positive feedback from the open-related. And something that we're truly proud of is that we also newly introduced a large model, we call it audit model. This auto model can help to distinguish the coordination as we can to standard sense based on the characteristics of the auto. And this is also a very good result that has been harvested from the [indiscernible].
Well, regarding the content creation, we have already received a very good result from artificial intelligence-generated models, no matter from the user scale or from the revenue of the users, we are foreseeing very good progress demand. In that way, it can actually facilitate the users of identifying new music and new songs and new content.
Another milestone I have to mention in Q1 of this year was we are working with Tencent AI and introduced the first music generation model for music. It is through this model, we're working with Shanghai National Orchestra to organize the ever-first AI in-commerce concept in China. Besides that, we are also keeping an eye on the seeding models in the industry, one is solar and another one is BIL, and we're exploring those new models to see how they can fit into our platform.