Alibaba Cloud

On September 24th, following the release of the Qwen3-2507 series, Aliyun announced today the launch of Qwen3-Max, the largest and most capable language model of the general team to date. The Qwen3-Max-Instruct official version further upgrades coding and intelligence capabilities, reaching industry-led standards in comprehensive benchmark tests that cover knowledge, reasoning, programming, command compliance, human preferences, intelligent mission and multilingual understanding. Qwen3-Max-Thi..

On September 23rd, another familiar late night, Aliun released and opened a new Qwen3-Omni, Qwen3-TTS and Qwen-Image-Edit-2509 for Google Nano Banana. Qwen3-Omni is the first original-to-end full-state AI model of the industry that can process multiple types of input for text, images, audio and video, and resolves the long-standing need for multi-modular models to move between different capabilities through text and natural voice real-time output..

In the early hours of the morning, Alithong released the next generation of basic model structures, Qwen3-Next, and opened up the Qwen3-Next-80B-A3B series based on the architecture. According to official sources, it is believed that Context Length Scaling and Total Parameter Scaling are two major trends in the development of large models in the future. In order to further enhance the training and reasoning efficiency of the model under long context and large-scale general parameters, a new Qwen3-Next model structure was designed. I don't know

August 27 news, yesterday evening, Ali Cloud announced the open source of a new multimodal video generation model Tongyi Wan phase Wan2.2-S2V, only a static picture and a piece of audio, you can generate a natural facial expression, consistent mouth shape, body movement silky smooth movie level digital human video. According to reports, the length of the video generated by this model can be up to minute level, which significantly improves the efficiency of video creation in the industries of digital human live broadcasting, movie and TV production, and AI education. Currently, Wan2.2-S2V can drive real people, cartoons, animals, digital people and other types of images, and supports portraits, half-body and full-body any...

August 22 news, Aliyun announced that Aliyun, super star Erya, in collaboration with Peking University, Nanjing University, Fudan University, Shanghai Jiaotong University, Zhejiang University, five universities and teachers, jointly launched the AI general public welfare series of courses "hands-on learning AI: general knowledge and practice of artificial intelligence," will be officially opened on September 1 for all colleges and universities across the country, all the majors of the teachers and students. Relying on Aliyun's cloud computing and AI technology, the AI general course creates a new teaching mode of "learning-practicing-evaluating-managing" theory + experiments, and builds a teaching mode that integrates theory and practice in depth, which is divided into theoretical courses...

August 9, AliCloud Tongyi Qianqian officially announced that AI programming intelligence body Qwen Code users can get 2000 (request) running privileges for free every day, and there is no Token limit. Its GitHub home page shows that 2000 ModelScope API calls per day are limited to users in mainland China (1000 OpenRouter API calls overseas). In addition, the official GitHub also provides developers with resources for accessing and using GitHub. Qwen Co...

July 23 news, AliCloud Tongyi team today released a notice announcing the official open source of Qwen3-Coder, which is called "intelligent programming engine for the world". This is also the team's most Agent-capable code model to date, in Agentic Coding, Agentic Browser-Use and Foundational Coding Tasks have achieved the SOTA effect of the open-source model, comparable to Claude Sonnet4. According to the presentation...

July 22, 2011 - AliCloud today updated its flagship Qwen3 model with an updated version of Qwen3-235B-A22B-FP8 Non-thinking mode (Non-thinking) named Qwen3-235B-A22B-Instruct-2507-FP8. AliCloud said that after communicating with the community and deep consideration , it decided to stop using the hybrid thinking model and switch to training Instruct and Thinking models separately to get the best quality. According to the introduction, the new Q...

June 9 news, domestic AI startup SiliconFlow (SiliconFlow) announced today that it has completed hundreds of millions of yuan in Series A financing. This round of financing was led by AliCloud, followed by old shareholders Innovation Works and other organizations, with China Renaissance acting as the exclusive financial advisor. Yuan Jinhui, founder of Silicon Mobility, said, "As a professional player who has been deeply engaged in the field of AI infrastructure for many years, Silicon Mobility has effectively solved a number of major problems in the development of the domestic AI industry through continuous technological breakthroughs and product innovation. This year, with Alibaba Tongyi Qianqian Qwen, De...

April 7 news, Aliyun official article today announced that recently, the Chinese Academy of Sciences Tibetan Plateau Research Institute jointly with Aliyun released a large model of multimodal reasoning "Luoshu" of water energy grain. According to the introduction, "Luoshu" large model integrates the scientific model "SiYuan" (Hydro Trace) and a thousand questions inference model QwQ-32B and multimodal model Qwen2.5-VL, can be a specific region in different time scales of the The system can accurately analyze and predict the amount and source of water coming to a specific region at different time scales, thus helping to regulate the dynamic balance of water supply, power generation and food production demand in water resources management. Currently, the "Luoshu" ...

April 1 news, Aliyun today said in an article, following the big model access to astronomical telescopes, the National Astronomical Observatory once again united with Aliyun to release the first international solar big model -- "Jinwu". It is said that the model is based on the Tongyi Qianqian series of open source models, and the current accuracy of the M5 level solar flare prediction exceeds 91%, the highest level of solar prediction at this level. Note: Solar flares are usually categorized into five levels: A, B, C, M, and X. When a solar flare erupts, a large number of charged high-energy particles attack the Earth's electromagnetic environment, which can cause widespread power outages around the globe, destroy on-orbit satellites, and threaten the safety of space stations in severe cases....

On March 31, 1AI learned from the official WeChat public number of Aliyun that recently, Nanjing University and Aliyun announced the launch of the artificial intelligence talent training cooperation program, and jointly cultivate AI innovation talents who are adaptable to future technological changes and have interdisciplinary thinking. Based on Aliyun's technological advantages in the field of cloud computing and AI big models and Nanjing University's disciplinary advantages in the field of artificial intelligence, the two sides will jointly design a general course on artificial intelligence that is both forward-looking and applied, develop a practical course on the application of big models that is interdisciplinary and integrated, and accelerate the "1+X+Y" three-level program at Nanjing University. The "1+X+Y" three-level "Artificial Intelligence General Knowledge Core Curriculum ...

March 21, 2011 - Aliyun today announced that it has entered into a partnership with the National University of Singapore (NUS), Thailand's Digital Economy Promotion Authority (DEPA), and 10 other universities and government agencies to launch AI and cloud computing training courses covering seven major Asian countries and regions, including Thailand, Singapore, and Hong Kong, China. Aliyun will build a Skills Center with the Institute of Operations Research and Analytics of the National University of Singapore (NUS) and the Hong Mei International Management Institute (HMI) in Singapore, which will provide cutting-edge training courses in the fields of big data and AI for students and professionals, and cultivate digital talents for the local community in Singapore. The National University of Singapore ...

March 6, 2011 - Research has shown that reinforcement learning can significantly improve the inference capabilities of models such as DeepSeek-R1, which achieves state-of-the-art performance by integrating cold-start data and multi-stage training, enabling deep thinking and complex inference. AliCloud Tongyi Qianqi officials today announced the availability of their latest inference model, QwQ-32B, a model with 32 billion parameters that rivals the performance of DeepSeek-R1 with 671 billion parameters, 37 billion of which are activated. This achievement highlights the importance of bringing reinforcement learning to ...

Aliyun announced on February 25th that its visual generation base model Wan Phase 2.1 (Wan) is open source. The open source adopts the loosest Apache2.0 protocol, and all the inference codes and weights of 14B and 1.3B parameter specifications are open source, while supporting text-generated video and graph-generated video tasks, which can be downloaded and experienced by global developers in Github, HuggingFace and Magic Ride communities. According to the introduction, the 14B Universal Phase model has outstanding performance in instruction following, complex motion generation, physical modeling, text-video generation, etc., and has been evaluated in the review set VBe...

Feifei Li's research team trained an AI inference model called s1 for less than $50 in the cloud, which performed similarly to cutting-edge inference models such as OpenAI's o1 and DeepSeek's R1 in tests of mathematical and coding ability. However, it was soon revealed that the s1 model was "not trained from scratch" and that its base model was the "Ali Tongyi Qwen" model. In this regard, Sina Technology asked AliCloud for confirmation, AliCloud confirmed the news. AliCloud responded: "They use Ali Tongyi Qwe...

January 29, 2011 - On the occasion of the New Year, AliCloud announced its new Qwen 2.5-Max hyperscale MoE model, which can be accessed by way of APIs, or you can log in to Qwen Chat to experience it, for example, by talking to the model directly, or using artifacts, search and other functions. According to the introduction, Qwen 2.5-Max uses over 20 trillion tokens of pre-training data and a well-designed post-training program for training. Performance AliCloud directly compares ...

Aliyun official announced on January 8, Tongyi Ling code AI programmer fully online, while supporting VS Code, JetBrains IDEs, front-end and back-end development full coverage, claiming that "the whole dialogue collaboration to achieve from 0 to 1 to complete complex coding tasks. The official said, from 0 to 1 development of a user login, data management features and beautification of the front-end page, for example, the past purely manual development needs to be front-end and back-end engineers to work with, at least half a day to spend time, now using Tongyi Ling code AI programmer, a programmer only needs 10 minutes...

January 7 news, 1AI from the official WeChat public number of Ali Cloud was informed that Transn Holdings (hereinafter referred to as Transn) has reached a cooperation with Ali Cloud, Tongyi Qianqian big model has been carried in the AI cell phone PHANTOM V Fold2 launched by TECNO, the technology brand of Transn, to create a deep localization of the "practical AI". "AI Through the specific AI buttons on the body, users can call the end-side large model with one key, and can still run multiple rounds of AI conversations smoothly in an offline environment, and realize document and call summaries. According to reports, Transn and Aliyun on the phone...

January 3 news, consumer AR brand ray bird innovation RayNeo and aliyun held a strategic signing ceremony on January 2, the two sides announced that in the field of AI eyeglasses to reach an exclusive strategic cooperation, Tongyi series of large models will provide exclusive customized technical support for the AI eyeglasses of ray bird innovation. It is reported that this is the first domestic large model and AI glasses hardware manufacturers to carry out in-depth customization cooperation, Aliyun Tongyi series of large models will provide a full set of AI technology and product support for Thunderbird's AI glasses products. The two sides said that the AI glasses will be the core carrier,...

January 2 news, ali cloud and big model enterprise zero one thing today reached model platform business strategic cooperation, the two sides will set up "industry big model joint lab", join hands to accelerate the big model from technology to application landing, further expand the ecological integration of industry big model. It is understood that the depth of strategic cooperation between the two sides of the industrial big model joint laboratory contains technology, business, talent and other sectors, the two sides will combine the two teams of big model R & D strength, the formation of the next generation from the base model technology exploration to industrial landing big model service combo, comprehensive through the AliCloud hundred refined big model platform model service level to the market. ...

December 18th, Intelligent Emergence today cited "multiple independent sources" as saying that Ali's AI application "Tongyi" was recently spun off from AliCloud and merged into Ali's Intelligent Information Business Group. Sources familiar with the matter said that the adjustment includes the product managers of Tongyi's To C direction, as well as the related engineering team, all of which have been adjusted to the Ali Intelligent Information Business Group. After the adjustment, Tongyi PC and App team and intelligent search product "Quark" level, the original Tongyi Labs, still remain in the Ali Cloud system. According to reports, Ali's big model family "Tongyi" previously included...

The AliCloud Tongyi Wanxiang team this week launched a new image editing model, ACE, which allows users to generate images and modify them with a single sentence. The tool can be applied to stylized portraits, split-screen production, interior design and other scenarios. The model has been open-sourced for developers around the world, and its localized stylization function has been launched on the Tongyi App. According to the official introduction, the ACE model can support both text-generated drawings and image editing, and the user can input spoken commands to complete tasks such as controllable visual editing, element editing, area redraw editing, and layered editing (stylizing specific objects in the drawings), which can be achieved by using only conversations. ...

On September 21, 2024, Alibaba announced a number of new developments in technological innovation and business development at the Hangzhou Yunqi Conference. Among them, the Magic Ride community officially launched the AIGC zone, aiming to provide developers with a comprehensive AI creation and development platform. The platform currently opens all function boards and GPU arithmetic for free, and the first batch of 157 selected multimodal models are on the shelves, including popular models in the community and a variety of stylized LoRa models contributed by designers. In the field of security, AliCloud announced a full upgrade of its cloud-native security capabilities, releasing for the first time its cloud-native network detection and response product, NDR, and at the same time promising...