Kuaishou's KwaiYii LLM, China's top short video platform operator, “significantly outperforms GPT-3.5 in overall performance and approaches the level of GPT-4 in certain metrics,” the company's founder said. said Cheng Yixiao, CEO and CEO. Wednesday's post-earnings conference call.
01:45
AI-generated cartoon series to be aired on Chinese state TV
AI-generated cartoon series to be aired on Chinese state TV
GPT-4 is OpenAI's most advanced LLM that is publicly accessible.
Kuaishou is confident that KwaiYii can rival GPT-4 within the next six months, Cheng said. The company also said KeTu's text-to-image model outperformed Midjourney V5, an AI image generation service launched a year ago by San Francisco-based Midjourney.
Meanwhile, Tencent said that its Hunyuan LLM is currently one of the best companies in the world.
The social media and video gaming giant is enhancing the ability of its AI models to create images and videos based on text prompts, company president Martin Lau Chi-ping said in a media interview on Wednesday. Speaking on a conference call, he called the effort “a most important undertaking.” This year is important for us.”
The Shenzhen-based company recently introduced an image-to-video model through partnerships with academic partners such as the Hong Kong University of Science and Technology and Tsinghua University in Beijing.
Dubbed Follow-Your-Click, the tool lets you animate any part of an image with a simple text prompt and a mouse click.
Both Tencent and Kuaishou are exploring ways to apply cutting-edge AI technology to their businesses.
Tencent's Lau said the Hunyuan model is being integrated into some of the company's enterprise products. This model is used by employees not only to improve ads to drive clicks, but also for day-to-day tasks such as coding and content generation.
ByteDance's Douyin's main rival, Kuaishou, is using KwaiYii to improve automated customer service in several of its businesses, including online human resources and real estate sales platforms, according to CEO Chen. It is said that it is using
He also said the company is looking to use AI to develop generative models that will serve as creative tools for users to reduce content creation barriers, improve video quality, and increase production efficiency. Ta.
“Sora's faster-than-expected launch presents tremendous opportunity for the short video ecosystem,” Cheng said.
