Today, Tencent officially announced that its AI assistant "Tencent Yuanbao" has launched a new image-to-text feature. Users can simply upload any image on the Tencent Yuanbao platform and leverage the advanced DeepSeek technology to enable AI to deeply understand and interpret the content of the image.

According to detailed explanations from Tencent officials, DeepSeek technology was previously mainly focused on text recognition and extraction from images. However, with continuous technological advancements, Tencent successfully integrated the multimodal understanding technology of HUNYUAN with DeepSeek. This allows Tencent Yuanbao to conduct in-depth analysis and precise interpretation of user-uploaded images without needing online searches. Users just need to send an image along with their questions, and Tencent Yuanbao will quickly provide insightful answers.

Tencent announced on February 17th that its AI assistant "Tencent Yuanbao" now fully supports free use of the full version of the DeepSeek-R1 model. Additionally, the HUNYUAN deep thinking model T1 (Thinker) is also available for users to test on the Tencent Yuanbao platform. Users can choose different models according to their needs to solve complex problems, thereby further improving work efficiency and satisfaction.

Compared to traditional generative models (often referred to as fast-thinking models), inference models like HUNYUAN T1 and DeepSeek R1 (slow-thinking models) have demonstrated superior performance in solving complex problems. They can analyze issues more thoroughly and provide more accurate and comprehensive answers.