ENLatest AI NewsOpenAI Criticizes AI Evaluation Benchmark: 731 Questions, Nearly a Third Have Flaws. 8-Month Passing Rate Rises from 23% to 80%, Now IneffectiveOpenAI publicly questioned the SWE-Bench Pro benchmark, pointing out that about 30% of its 731 test tasks have evaluation flaws. The benchmark, launched by Scale AI, is an industry authority for measuring large model programming capabilities. However, OpenAI warned that the passing rate of cutting-edge models has surged from 23.3% to 80.3% within 8 months, which is unusually fast, indicating doubts about the reliability of the evaluation.7 hours ago59.2KAI Pay-as-You-Go Pricing Deters Executive Leaders: Nearly One-Third Admit to Not Understanding the Economic Account, Computing Bills Have Become a Black BoxA KPMG survey of 2,145 executives in 20 countries reveals unease over AI’s new pay-per-use pricing. The past model of fixed-price contracts with AI companies subsidizing computing costs is unsustainable as prices surge, forcing a defensive shift and ending cheap AI access…..8 hours ago12.6KAmazon Exposed to Develop AI Agent Moonraker: Alexa Will Support Multi-step Complex Task SequencingAmazon is secretly developing an AI agent project “Moonraker” as an extension of Alexa Plus. This technology breaks the limitation of single instructions, deeply understands compound requests, and automatically executes multi-step operations in logical order, such as completing tandem calls with one command…..8 hours ago9.7KA regular camera can navigate autonomously: Mistral releases the 8B model Robostral Navigate, performance surpasses multi-camera solutionsThe French AI company Mistral has launched a lightweight robot navigation model called Robostral Navigate with only 8B parameters. The model can achieve complete autonomous navigation in complex environments using only a regular RGB camera, without the need for lidar or depth sensors, and is suitable for various indoor and outdoor scenarios, significantly reducing hardware costs and deployment complexity.8 hours ago11.2KStep Star’s First AI Intelligent Phone Will Be Released, Leading Ahead of OpenAIStep Star announced a new AI agent device, launching its AI terminal brand, agent system, and first AI agent phone. It becomes the first native agent phone from a major model company, with market release ahead of OpenAI’s plan. Manufactured by Huaqin Technology, the two have formed a deep partnership…..9 hours ago12.5KMeta Exposed to Be Developing All-Day AI Glasses: Support for Super Perception and Seamless Audio and Video CaptureMeta is developing an all-day multimodal sensing smart glass prototype with ‘super perception,’ automatically recording audio and photos every few seconds. Users can ask Meta AI about captured content instantly. For privacy, raw media isn’t stored; only metadata is uploaded…..9 hours ago10.6KWorld’s First AI Actress-Starring Film Begins Production: No Body, No Childhood, but Can Access All of Human MemoryUK-based Particle6 begins shooting Misaligned, starring AI actress Tilly Norwood in what it calls the first feature film with an AI lead. The coming-of-age comedy-drama unfolds in a surreal digital realm, the Tillyverse, where the AI protagonist—bereft of body, childhood, or past—grapples with identity…..9 hours ago9.7KCollege Entrance Exam Admission Begins: Qwen Integrates with EMS to Launch Real-Time Tracking and Reminders for Admission LettersQianwen Gaokao and China Post EMS launched a delivery tracking service for admission letters, pushing authorization reminders via a voluntary calendar, proactively notifying users upon shipment or logistics updates, allowing real-time tracking…..10 hours ago12.1KXiaomi’s New AI Strategy: Building the Intelligent Infrastructure of the Future, Reshaping the Industry Landscape!Xiaomi reorganizes its XiaoAi technical team to address the AI large model trend, integrating AI deeply into all businesses. This shift marks a transition from previous single-product development to a strategy driven by large models, demonstrating foresight and potentially building future competitive advantage…..12 hours ago16.1KHe Gao AI Sector Remains Hot: Zhipu and MiniMax Show Strong Performance on the Unblocking DayHong Kong’s AI sector rallied with Zhipu and MiniMax in focus. Despite lockup expiry concerns, shares surged, showing market confidence in AI assets. Zhipu led with an intraday gain of over 20%, driving the sector after lockup fears subsided…..12 hours ago10.7KPrime Intellect Completes $130 Million Funding, Valuation Reaches $1 BillionPrime Intellect raised $130M Series A, valued over $1B, becoming a unicorn. Led by Radical Ventures, with Nvidia, Intel, Dell VC, Iconiq, and Perplexity founder participating…..12 hours ago9.9KAI Coding Circle Sees Another Major Development: Swedish Star Lovable’s Valuation May Double to $13.2 BillionSwedish AI programming company Lovable is currently raising a new round of $300 million in funding, with i