Search AI Products and News

In the field of artificial intelligence, reward models are a critical component for aligning large language models (LLMs) with human feedback, but existing models face the issue of "reward hacking." These models often focus on superficial features, such as the length or format of responses, rather than identifying genuine quality metrics, such as factual accuracy and relevance. The root cause lies in standard training objectives failing to distinguish between spurious associations and true causal drivers present in the training data. This failure leads to fragile reward models (RMs), which generate misaligned policies.

Zhipu AI officially open-sources its latest general vision model, GLM-4.1V-Thinking, based on the GLM-4V architecture, which introduces a chain-of-thought reasoning mechanism, significantly enhancing its capabilities for complex cognitive tasks. The model supports multimodal inputs such as images, videos, and documents, and excels in diverse scenarios including long video understanding, image question answering, subject problem-solving, text recognition, document interpretation, grounding, GUI Agent, and code generation, covering a wide range of industry application needs. GLM-4.1V-9B-Thinking

Welcome to the AIbase [AI Daily Report] section! Spend three minutes a day to learn about the latest AI events, helping you understand AI industry trends and innovative AI product applications. For more AI news, visit: https://www.aibase.com/zh1. Baidu officially releases the WENXIN Large Model 4.5 series and fully opens it to the public, featuring ten new models with various parameter configurations. These models are trained and inferred using the PaddlePaddle framework, achieving a FLOPs utilization rate of 47%, and perform well in multi-modal text tasks.

Recently, news about OpenAI's upcoming release of GPT-5 has attracted widespread attention in the technology industry. According to insiders, GPT-5 has already started a gradual test and is expected to be officially launched in July this year. This new model will adopt a multimodal design, meaning it can not only process text input but also understand speech, images, code, and even videos, completely changing the way we interact with AI. Sam Altman, CEO of OpenAI, stated that the launch of GPT-5 will mark a new era in AI.

Recently, the Qwen VLo multimodal large model was officially released, achieving significant advancements in image content understanding and generation, offering users a brand-new visual creation experience. According to the introduction, Qwen VLo has been comprehensively upgraded based on the advantages of the original Qwen-VL series models. The model not only can accurately understand the "world", but also can perform high-quality re-creation based on understanding, truly achieving the transition from perception to generation. Users can now access Qwen Chat (chat.qwen.ai)

Recently, a research team from the University of Hong Kong, The Chinese University of Hong Kong, and SenseTime has released a groundbreaking framework - GoT-R1. This new multimodal large model significantly enhances the semantic and spatial reasoning capabilities of AI in visual generation tasks by introducing reinforcement learning (RL), successfully generating high-fidelity and semantically consistent images from complex text prompts. This advancement marks another leap in image generation technology. Currently, although existing multimodal large models have made significant progress in generating images based on text prompts

The new "File" tab in Grok Web, developed by xAI, is set to launch, offering users an all-in-one file management experience. This feature will integrate various file types such as images, spreadsheets, text, and code, significantly improving work efficiency and convenience. The new function allows users to browse, create, upload, and edit files within a single unified interface. Whether handling complex code, editing text documents, or managing images and spreadsheets, users can easily perform these tasks through the "File" tab. This design simplifies the file management process for those working with multiple file types.

Recently, at Baidu's AI Open Day, Wenxin Kaima, Baidu's intelligent code assistant, achieved a major breakthrough. Its standalone AI-native development environment tool, Comate AI IDE, was officially launched. This industry's first multimodal, multi-agent collaborative AI IDE not only introduces the groundbreaking feature of converting design drafts into code with one click, but also provides efficient, intelligent, and secure programming experiences for domestic enterprises and developers.

With the continuous progress of artificial intelligence technology, multimodal data processing has gradually become a popular topic. Recently, the globally renowned electrical appliances brand Panasonic launched its latest R&D multimodal large model — OmniFlow. This model can efficiently convert between multiple modalities such as text, image, and audio, achieving any-to-any generation tasks, providing users with a more flexible experience. The design concept of OmniFlow is based on modularity, allowing the various components of the model to be independently pre-trained. This approach not only improves training efficiency but also avoids...

Recently, an open-source RAG (Retrieval-Augmented Generation) engine called RAGFlow has attracted significant attention in the industry. This enterprise-level AI tool, based on deep document understanding, provides companies with a brand-new solution for processing complex documents and achieving precise question answering through its powerful multimodal data processing capabilities and efficient workflows. RAGFlow: The Pioneer of Deep Document Understanding RAGFlow is a fully open-source RAG engine that focuses on deep document understanding, aiming to help businesses and individuals extract valuable information from massive amounts of unstructured data.

The Nanometer AI Super Search Intelligent Entity under the 360 Company has undergone a major update, adding new features such as multimodal content generation, cross-domain professional search, and smarter task preview functions. From one-click generation of PPTs, PDF reports to automatically integrating videos, voiceover scripts, and storyboard planning, Nano AI redefines the boundary of AI search and creation with more efficient and intuitive experiences. AIbase comprehensively organizes the latest social media dynamics to help you deeply understand the latest breakthroughs of Nano AI. Multimodal Generation: Handle everything from PPTs to videos with one click.

The Meta AI research team has made another breakthrough in the artificial intelligence field, officially releasing the new video understanding model V-JEPA2 (Video Joint Embedding Predictive Architecture2) on June 11, 2025. Led by Meta's chief AI scientist Yann LeCun, this model has opened up new possibilities in video understanding and physical world modeling with its innovative self-supervised learning technology and zero-shot robot control capabilities.

In recent years, the field of artificial intelligence has undergone tremendous changes, especially with large language models (LLMs) making remarkable progress in multi-modal tasks. These models have demonstrated strong potential in their ability to understand and generate language, but most current multi-modal models still adopt auto-regressive (AR) architectures, which limit inference processes to be rather monotonous and lacking flexibility. To address this limitation, a research team from The University of Hong Kong and Huawei Noah's Ark Lab has proposed a brand new model – FUDOKI, aiming to break these constraints. The core innovation of FUDOKI is

Building on the GPT-4o released last year, OpenAI has made significant updates to its advanced voice mode, making voice communication more natural and aligned with human conversation styles. This advanced feature is based on the native multimodal model, which can quickly respond to audio inputs, with the fastest response time at 232 milliseconds and an average response time of 320 milliseconds, nearly matching the speed of human conversations. At the beginning of this year, OpenAI had already made minor updates to this voice mode, improving interruption frequency and accent handling.