Search AI Products and News

Researchers from Hong Kong and the UK have recently proposed a novel image tokenization method designed to convert images into digital representations (tokens) in a more compact and accurate manner. Unlike traditional methods that distribute information uniformly across all tokens, this method employs a hierarchical structure, capturing visual information layer by layer, thereby improving the quality and efficiency of image reconstruction. Traditional image tokenization techniques typically divide each part of an image equally into multiple tokens, whereas the new method adopts a hierarchical structure. Initial tokens encode rough shapes and structural elements,

To address the increasingly competitive AI market, OpenAI recently introduced a new API called Flex Processing. This new option allows users to utilize AI models at a lower cost, although with some trade-offs in response speed and availability. Flex Processing is designed to support lower-priority and non-production tasks such as model evaluation, data enrichment, and asynchronous workloads. Specifically, Flex Processing offers significantly reduced costs. For example, when using the o3 model, Flex Processing's price is...

OpenAI recently released its latest reasoning models, o3 and o4-mini, marking a significant breakthrough in the field of artificial intelligence. These models not only surpass previous versions in reasoning capabilities but also achieve image reasoning for the first time, integrating visual information directly into the thinking process. o3, hailed as a "genius-level" model, particularly excels in tasks such as programming and mathematics, achieving an accuracy rate of 87.5%. The newly released o3 and o4-mini models demonstrate exceptional performance in multimodal processing, possessing...

Zhihu AI, a leading player in the Chinese large language model market, has officially begun its initial public offering (IPO) process, marking a significant milestone for the industry's 'Big Six' companies.

Artificial intelligence is rapidly permeating everyday productivity tools. Recently, AIbase learned from social media that Google Sheets has officially launched an AI function, providing users with efficient bulk data processing capabilities through the built-in =AI() function. While this feature is currently in alpha testing and available to select users, social media users have already expressed amazement at its performance. Below is AIbase's in-depth report on this update, analyzing its technical highlights and potential impact. Google Sheets

Los modelos de lenguaje visual de código abierto Kimi-VL y Kimi-VL-Thinking han superado a GPT-4o en varias pruebas de referencia. Estos modelos representan un avance significativo en el campo de la inteligencia artificial, combinando la capacidad de procesamiento del lenguaje natural con la comprensión de imágenes.

In education, we're taught to "show your work." Now, advanced AI models claim to do just that. However, new research reveals that these models sometimes obfuscate their true reasoning processes, fabricating elaborate explanations instead. A recent study from Anthropic's research team, investigating simulated reasoning (SR) models including their own Claude models and DeepSeek's R1, found these models often misrepresent their 'thinking' when

IBM on Monday launched its latest mainframe hardware, the IBM z17. This fully encrypted mainframe, powered by the IBM Telum II processor, is designed for over 250 AI use cases, including AI agents and generative AI applications. While mainframes may be considered legacy technology by some, 71% of Fortune 500 companies still use them, according to sources. According to market research firm Market Research Future, by 2024...

Tencent has recently unveiled GeometryCrafter, a new AI model released via Hugging Face. This model excels at consistent geometric estimation in open-world videos, quickly gaining attention in the tech community. Leveraging diffusion priors, GeometryCrafter opens up new possibilities for deep understanding and processing of video content, offering benefits to creators and researchers alike.

Yuanbao, Tencent's AI assistant, has always been committed to enhancing user efficiency through powerful model capabilities. This update allows users to directly upload common document formats including Word, PDF, and Excel, as well as development files such as .py, .java, and .json. Yuanbao can not only understand the content of these files but also interpret code based on user needs.

Tencent Yuanbao has announced a new feature designed to simplify the understanding and analysis of Excel spreadsheets. This upgrade allows users to ask questions in natural language, and Yuanbao will directly extract data, perform calculations, and provide answers from the spreadsheet, even highlighting key information. This significantly streamlines complex spreadsheet operations.