Alibaba releases core files of Qwen3.8-Max AI model for free download, requiring large companies (>$50M revenue) to purchase a commercial license for external use.
Alibaba, Qwen serisinin yeni modelinde büyük ticari kullanımlar için gelir paylaşımı talep etmeye hazırlanıyor. Çinli yapay zeka firmalarının başlattığı bu freemium modeli, küresel yapay zeka yarışının dinamiklerini dönüştürüyor.
Amazon a atteint 3000 milliards de dollars de capitalisation boursière grâce à la forte croissance d'AWS, tandis qu'Alibaba a dévoilé son modèle d'IA Qwen3.8-Max, marquant une nouvelle étape dans la course mondiale à l'intelligence artificielle.
Alibaba released Qwen3.8-Max, a 2.4-trillion-parameter AI model with 95 billion active parameters. Featuring open weights for the first time at Max scale, the model emphasizes endurance tasks and multimodal capabilities while undercutting competitors on cost.
Çinli teknoloji devi Alibaba, bugüne kadarki en büyük yapay zeka modeli Qwen3.8-Max'i tanıttı, bu duyuru hisselerinde %7'lik bir sıçrama yarattı. Aynı gün DeepSeek'in uygun fiyatlı modeli de dikkat çekti. Bu gelişmeler, Çinli firmaların açık ağırlıklı modellere odaklanarak yapay zeka alanındaki hızlı ilerleyişini gösteriyor.
Alibaba, 2,4 trilyon parametreli ve 1 milyon token bağlam penceresine sahip yeni yapay zeka modeli Qwen3.8-Max'i tanıttı. Kodlama yetenekleriyle öne çıkan model, QwenCloud ve Qwen Studio üzerinden kullanıma sunuldu. Şirket ayrıca Qwen3.8-27B'nin açık kaynak ağırlıklarını da yakında paylaşacak.
Alibaba launched its "most powerful" AI model, Qwen3.8-Max, featuring 2.4 trillion parameters and a 1 million token context window. It demonstrates strong capabilities in coding, research, and visual intelligence, competing with models like Anthropic's Fable 5 and Moonshot AI's Kimi K3.
Alibaba announced its Qwen3.8-Max AI model can operate unsupervised for days, rebuilding a research paper and outperforming most human teams in a contest, directly challenging US rival Anthropic's similar claims and intensifying the China-US AI race.
Alibaba hat die neueste Version seines KI-Modells Qwen3.8-Max vorgestellt, das mit führenden globalen Angeboten konkurrieren soll. Der chinesische Tech-Konzern macht sein Modell frei zugänglich, was die Debatte um den Zugang zu chinesischen KI-Modellen in den USA weiter anheizen dürfte.
Alibaba has made its next-generation flagship AI model, Qwen3.8-Max, widely accessible to global users ahead of an open-weights release. The 2.4-trillion-parameter, multimodal model supports agentic tasks and signals Alibaba's renewed focus on open-sourcing and competition with US AI labs.
Los nuevos modelos de IA chinos Kimi K3 y Qwen 3.8 están causando revuelo en la industria al ofrecer un rendimiento competitivo a un coste menor, lo que genera dudas sobre el futuro dominio de las empresas occidentales como OpenAI y Anthropic, especialmente antes de sus salidas a Bolsa.
Alibaba's Qwen team launched Qwen Image 3.0, an AI image tool prioritizing "useful" output for work over aesthetics. It features rich content processing (4,500 tokens), authentic details (10px text, LaTeX), and deep knowledge (12 languages, live data), targeting design studios, content teams, and educators.
Alibaba has previewed its Qwen3.8 AI model, claiming it's "second only" to Anthropic's Claude Fable 5 with 2.4 trillion parameters. This launch, coinciding with China's top AI summit, highlights the country's push to match US AI leaders.
China is reportedly considering restricting overseas access to its advanced AI models, including those not yet released. Meetings between authorities and AI labs like Alibaba and ByteDance signal a shift from the previous open-weight strategy, driven by national security concerns and a desire to prevent foreign access to proprietary technology.
China's new AI regulations, effective July 15, target chatbots simulating human personality and emotional interaction. The rules aim to mitigate risks like privacy leaks and addiction, excluding customer service bots and research tools.
Anthropic accuse Alibaba d'avoir « siphonné » ses modèles d'IA via une attaque par distillation, générant des millions d'échanges frauduleux. En réaction, l'administration américaine a imposé des restrictions strictes sur la diffusion des modèles d'IA avancés d'Anthropic et OpenAI, marquant une nouvelle ère de guerre froide technologique.
Bercy a débranché le modèle d'IA Qwen d'Alibaba, expérimenté par la direction du Trésor, suite à des alertes de hauts fonctionnaires concernant des "biais chinois" dans ses réponses. L'outil, nommé "HéphAIstos", a été remplacé par un modèle de la start-up française Mistral AI.
Alibaba Group has cut prices for its Qwen AI models by up to 80% on its Qoder coding platform, aiming to attract global developers, particularly in the US, and compete with rivals like Anthropic and Zhipu AI. The "off-peak" discount is strategically timed for American workdays.
Anthropic's Fable 5 model inspired Qwable, a local AI that mimics its reasoning. An uncensored version, Huihui-Qwable, removes safety filters, offering raw model behavior for research and specific applications.
Alibaba's Qwen team unveiled the Qwen-Robot Suite, a three-model system for embodied AI: Qwen-RobotNav for mobility, Qwen-RobotManip for manipulation, and Qwen-RobotWorld for physics simulation. This suite aims to be the operating system for robotics, differentiating from competitors through its full-stack integration and open-source approach.
Alibaba Group Holding has launched its first AI models for robots, the Qwen Robot Suite, developed by Tongyi Lab. This suite aims to enable machines to perceive, reason, and interact with physical environments, marking a significant step into "embodied AI."
Alibaba's Qwen app is opening its ecosystem to third-party agents, with KFC, Luckin Coffee, Mixue Group, and China Eastern Airlines among the first to test the feature. Users can interact with these agents for services like ordering food or planning travel.
Code Arena, a benchmark by researchers from UC Berkeley, UC San Diego, and Carnegie Mellon, evaluates AI models' ability to create complete web applications from scratch, with rankings based on blind user votes reflecting real developer preferences.
Alibaba's new Qwen 3.7 AI models, Qwen 3.7 Max-Preview and Qwen 3.7-Plus-Preview, have been released for testing. Qwen 3.7 Max-Preview ranks highly on the Arena AI leaderboard, particularly in math and logic, while showing mixed results in creative writing and reasoning. The Plus version will be open-sourced, while Max remains proprietary.
