SqueezeBits Launches 'Yetter' AI API Open Beta: 2.5x Faster, Half the Cost

## Opening New Horizons for Generative AI: SqueezeBits Unveils ‘Yetter’, a High-Speed, Low-Cost API Service AI optimization specialist startup SqueezeBits is drawing market attention by launching the open beta of its innovative API service 'Yetter,' designed to lower the barrier to utilizing generative AI. Yetter, which enables easy utilization of generative AI's powerful features via API without the need for expensive infrastructure setup, was born from proprietary inference engine technology and extensive hardware support. SqueezeBits announced that Yetter boasts performance that surpasses existing global competing services. Specifically, AI inference speed is approximately 2.5 times faster, and service costs can be significantly reduced by up to half, earning praise for paving the way for companies to more efficiently adopt and expand generative AI. This is expected to become an attractive alternative for many companies that have hesitated to adopt generative AI due to cost and speed concerns. These innovative achievements stem from SqueezeBits' proprietary technology. Their self-developed inference engine simultaneously achieves AI model lightweighting and performance optimization, supporting a wide range of AI semiconductor environments from GPUs to the latest NPUs, enabling flexible service use unconstrained by specific hardware. This means customers can enjoy optimal performance in their desired environment. Currently, Yetter supports image generation from text, editing existing images, and video generation using both text and images. In the future, the plan is to gradually add more diverse AI models, such as text generation features based on Large Language Models (LLMs), to expand the scope of services. Since its establishment in 2022, SqueezeBits has focused on AI model lightweighting and inference optimization technologies, accumulating technical expertise through collaborations with global hardware companies such as Intel and Rebellions. CEO Kim Hyung-jun stated, "Yetter is the culmination of our accumulated AI lightweighting optimization technology and extensive understanding of hardware," adding, "Our ultimate goal is to maximize the potential of AI technology while providing realistic cost-efficiency, thereby helping all businesses achieve AI innovation."