Search AI Products and News

Tencent open-sourced 'Hunyuan-A13B', an 80B-parameter MoE model with dynamic inference (130B active params). Supports 256K context, trained on 20T tokens. Achieves 87.3% on AIME2024 math, but comparisons show version disparities.....

Recently, a research team from Renmin University, Shanghai Artificial Intelligence Laboratory, University College London, and Dalian University of Technology revealed an important finding in the reasoning process of large models: when the model is thinking, the 'thinking words' it uses actually reflect a significant increase in its internal information. This research result provides a new perspective for better understanding the reasoning mechanisms of artificial intelligence through methods of information theory. You may have seen large models output some language that seems human-like when answering questions, such as "Hmm..." or "Let me think...".

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 [AI Daily] segment! This is your guide to exploring the world of artificial intelligence every day. Every day, we present you with the latest content in the AI field, focusing on developers, helping you understand technology trends and innovative AI product applications. Click to learn more about fresh AI products: https://top.aibase.com/1: Moon's Dark releases the first self-developed agent Kimi-Researcher. Moon's Dark launches Kimi-Researcher, which has strong multi-turn search and reasoning capabilities, and performs well in HLE testing.

Recently, Kimi intelligent assistant under the Moonshot Dark Side announced that its first agent product - Kimi-Researcher (Deep Research) has officially started small-scale gray-scale testing. This new generation of agent models trained based on end-to-end autonomous reinforcement learning (end-to-end agentic RL) technology aims to provide users with efficient and in-depth deep research services. Kimi-Researcher demonstrates strong autonomous planning and execution capabilities when facing problems. It can not only

At the theme speech of Huawei Developer Conference HDC2025 held this afternoon, Zhang Ping'an, Executive Director of Huawei and CEO of Huawei Cloud Computing, officially announced the release of PanGu 5.5, marking the comprehensive upgrade of five basic models in natural language processing (NLP), computer vision (CV), multimodal, prediction, and scientific computing.

This article provides an in-depth analysis of the innovative architecture and technical advantages of the Kimi-Dev-72B open source large model, exploring how it addresses efficiency pain points in software development through its three-stage training paradigm and environment-integrated reinforcement learning mechanism, featuring core functions such as automated issue localization and precise code repair.

A new study led by Apple researchers has poured cold water on the current large reasoning models (LRMs) that are highly anticipated. The research found that when solving complex tasks, reasoning models designed to simulate thinking processes, such as Claude3.7Thinking and Deepseek-R1, not only failed to show advantages but also exhibited serious problems like insufficient reasoning and performance collapse. This study tested four classic logical puzzles: the Tower of Hanoi, Checkers, River Crossing, and Block World. These puzzles allow precise control over the tasks.

NVIDIA officially released Llama Nemotron Nano VL on June 3, 2025. This compact visual-language model (VLM) is specially optimized for document intelligence processing. The model topped the OCRBench v2 benchmark test, showcasing its outstanding capabilities in handling complex documents, charts, and video frames. With efficient inference performance and flexible deployment options, Llama Nemotron Nano VL provides enterprises with solutions from the cloud.

Recently, Xiaomi Technology Co., Ltd. transferred multiple 'Kimi' trademarks to Beijing Yuelong Anmai Technology Co., Ltd. This trademark was first applied for in 2013 and is currently mostly registered. As a company that has always been active in trademark protection, Xiaomi has sued through legal means multiple times due to trademark infringement issues. This transfer may be part of its strategic adjustment to focus on core businesses, aiming to optimize non-core trademark resources. Beijing Yuelong Anmai Technology Co., Ltd. was established in April 2023 with a registered capital of 1 million RMB, and the legal representative is

Recently, the MiMo-VL multi-modal model developed by Xiaomi Company has taken over the baton from MiMo-7B, showcasing powerful capabilities in multiple fields. The model significantly outperforms the same-size benchmark multi-modal model Qwen2.5-VL-7B in tasks such as general question answering and understanding inference for images, videos, and language, and its performance in GUI Grounding tasks is comparable to that of specialized models, preparing for the arrival of the Agent era.

Recently, the Windows version of Sogou Input Method has undergone a major upgrade with the launch of the new AI Wanzai Quick Search function, providing users with a more efficient and convenient office and study experience. The AI Wanzai Quick Search function is characterized by its one-click trigger and fast response. Users no longer need to follow the traditional three-step search process of opening a browser👉🏻search engine👉🏻input content. They can directly type on any spot on the desktop or use the shortcut key Ctrl+Shift+Space”, which can summon the All-Purpose Quick Search” function within 0.5 seconds, allowing for faster information retrieval.