Latest news with #QwenVLo
Hindustan Times
01-07-2025
- Automotive
- Hindustan Times
Alibaba releases Qwen VLo, new image-generation tool to challenge Chatgpt-4o
Alibaba Group has released a new variation of its artificial-intelligence technology that will allow users to generate and modify images from text and visuals. The company released Qwen VLo as a part of a series of AI services provided by the company. This new model is an upgrade of the earlier Qwen2.5-VL and generates both text-to-image and image-to-image. Through a fascinating technology called progressive generation, the users will be able to see the process as an image is created. While giving instructions on Qwen VLo the user will be free to write in multiple languages, including in Chinese and English. (Reuters File Photo) In a post on X the company announced the release of new model along with its features and a link to access it. According to the blog post by the company on gitihub, Qwen VLo is a unified multimodal understanding and generation model. Not only does it 'understands the world' but it also generates high quality images based on that understanding. Text-to-image and image-to-image generation Through Qwen Vlo you can directly send the prompt like 'generate a picture of dog' or 'upload an image of a dog' and ask to make edits in the image. According to the blogpost, the previous models struggled with semantic inconsistencies like misinterpreting a car as another object or failing to retain key features of the car. With Qwen VLo the company fixed that and it can correctly identify the key features of a car like its model, color etc. Open ended instruction based editing While editing an image Qwen VLo will respond to open ended instructions like add a sun to the sky or make the photo look like 19th century. It even allows the user to make traditional perception tasks like predicting depth maps, segmentation maps, detection maps, and edge information. It can perform multiple of these editing functions at the same time. Multilingual support for prompt While giving instructions the user will be free to write in multiple languages, including in Chinese and English. According to the company, the model will understand instructions regardless of the language. Alibaba, popularly known for its e-commerce services, has been integrating AI and building standalone offerings around Qwen. In February, Chief Executive Officer Eddie Wu went so far as to say the company's 'primary objective' is now artificial general intelligence, meaning a goal to build AI systems with human-level intellectual capabilities.
Hindustan Times
01-07-2025
- Business
- Hindustan Times
Alibaba unveils Qwen VLo: Free AI model for text-to-image and photo editing
Alibaba has launched a new AI model for image generation called Qwen VLo. This model builds on the earlier Qwen 2.5 vision language system, and now offers enhanced features for creating and editing images. Qwen VLo supports both text-to-image and image-to-image generation and understands text prompts in several languages, including English and Chinese. Alibaba has launched Qwen VLo, a new AI model for generating and editing images using text prompts.(@Alibaba_Qwen) The Alibaba's Qwen team announced the release of Qwen VLo on their official social media handle, X. The model, formally known as Qwen3-235B-A22B, is now available for free to use through Alibaba's chat platform. Users do not need to log in to access it. This AI model will allow users to either generate images from text prompts or descriptions or modify existing images by providing written instructions. Also read: Super App race: How US startups are building all-in-one platforms One of the model's notable features is 'progressive generation,' which lets users see images form step-by-step. For example, users can ask the model to create an image of a robo dog or cat or upload a picture and request specific changes like adding an object to the image. The AI then applies the edits while maintaining the integrity of the original image. Qwen VLo AI Tool: Advanced Features According to Alibaba's GitHub page, Qwen VLo has improved capabilities in understanding images, which enables it to perform precise inline edits without distorting the input. It can also handle vague or open-ended prompts better, which will help it produce images more closely aligned with what users expect. Also read: AR in retail: How Augmented Reality is transforming shopping experiences Beyond generation and editing, the model can also perform image annotation tasks such as edge detection, segmentation, and prediction mapping. Future updates will allow Qwen VLo to process multiple images at once and combine them based on user instructions. On the other hand, text rendering within images has also improved, with the model accurately producing text in various fonts. Additionally, Qwen VLo supports dynamic input aspect ratios, including extreme dimensions like 4:1 and 1:3. Alibaba plans to introduce image generation in multiple aspect ratios soon. Also read: Self-Regulation in Action: How Probo's Safety Features Promote Responsible Gaming The development of Qwen VLo reflects Alibaba's growing focus on AI. The Hangzhou-based company, which is primarily known for its e-commerce operations, has been expanding its investment in AI research through the Qwen brand. Earlier this year, Alibaba CEO Eddie Wu emphasised the importance of developing artificial general intelligence (AGI), signalling the company's move towards creating AI systems with advanced reasoning capabilities. With Qwen VLo, Alibaba aims to compete with other AI developers both in China and internationally. The company faces competition from firms such as DeepSeek, which recently gained attention for releasing a powerful AI model developed with limited funding. Mobile finder: iPhone 16 LATEST price, specs and all details
International Business Times
30-06-2025
- Business
- International Business Times
What Is Qwen VLo? Alibaba's Launches Image Generation Model, to Compete with OpenAI and Google
Chinese technology companies have now openly entered in global AI competition and are launching robust and efficient AI models frequently to compete with US Silicon Valley's popular models. Alibaba, an e-commerce giant, has introduced its powerful new image-generation model, Qwen VLo. The advanced AI tool has been developed to create, transform, and comprehend visuals through text and image prompts and represents yet another huge milestone in China's aspirations to take the global lead in AI. Launched recently, Qwen VLo is building on the success of earlier models QwenVL and Qwen2. 5 VL, but with a substantial step up in quality and control. The new model employs a progressive approach to generation, constructing images incrementally from top to bottom and left to right. This approach offers greater accuracy, especially in detailed or complex visual edits. Qwen VLo can efficiently understand user guidelines with open-ended descriptions, like- changing artistic style, applying weather effects to a scene, or time period of a scene. It is capapble of taking instruction in multiple languages, including Chinese and English. Unlike previous designs, Qwen VLo is able to retain crucial details while making the changes. For instance, if a user wants to make only a color adjustment, it doesn't change parts of the image that are not directly affected. This also makes it great for professionals who need to design posters, illustrations, banners, and social media graphics. One of its impressive features is multiple image input, in which users can upload multiple images and instruct the AI to combine and modify them together. For example, placing products inside a basket based on user commands. Although this feature is not fully launched yet, it illustrates the capabilities of the model. Qwen VLo also features dynamic resolution training, where it can resize images into formats such as 1:1, 3:4 or 16:9. With all these impressive attributes, Alibaba advises that the model is still in the preview stage and that some errors or inconsistencies might occur. Alibaba, best known for e-commerce, is making a big bet on AI. Its CEO, Eddie Wu recently stated the company's future lies in developing human-level AI systems. In February, Alibaba announced that it would spend $52 billion over the next three years on AI-focused infrastructure.
Forbes
30-06-2025
- Business
- Forbes
China's AI Leap Forward: Tencent And Alibaba's New (And Faster) Models
Artificial Intelligence in China and how it competes with US In a bold stride toward global AI dominance, two of China's leading technology giants, Tencent and Alibaba, have unveiled groundbreaking AI models that signal both technological advancement and intensifying competition in the artificial intelligence landscape. Tencent's Hunyuan-A13B, an open-source hybrid reasoning model, and Alibaba's Qwen-VLo, a creative multimodal model akin to OpenAI's ChatGPT-4o, mark significant milestones in China's AI ecosystem. These releases not only showcase cutting-edge innovation but also pose strategic challenges for global competitors and opportunities for business executives navigating the AI-driven future. Tencent's Hunyuan-A13B: Efficiency Meets Power Tencent's Hunyuan-A13B is a Mixture-of-Experts (MoE) model with 80 billion total parameters, activating just 13 billion during inference, making it remarkably efficient. The model's performance rivals industry leaders like OpenAI's o1 and DeepSeek's R1. This efficiency reduces inference latency by 2.2-2.5x compared to larger models like Alibaba's Qwen3-A22B, according to posts on X and Tencent's technical reports. The open-source nature of Hunyuan-A13B, licensed, democratizes access for developers and small-to-medium enterprises, fostering rapid adoption and innovation. Its integration with frameworks like vLLM and TensorRT-LLM, ensures scalability across diverse applications, from coding to agentic tasks like tool-calling and data analysis. Alibaba's Qwen-VLo: Redefining Creative AI Alibaba's Qwen-VLo, part of the Qwen model family, pushes the boundaries of multimodal AI with its 'progressive generation' approach. This model integrates text, image, audio, and video processing, enabling dynamic workflows such as text-to-image generation, natural language-based image editing, and multilingual text creation across 119 languages. Unlike traditional models, Qwen-VLo supports multi-image input prompts and adapts to varying resolutions and aspect ratios, offering flexibility for creative industries. Its ability to 're-create' based on deep content understanding positions it as a competitor to OpenAI's ChatGPT-4o, with Alibaba claiming superior performance in video benchmarks. Qwen-VLo's open-source variants have garnered over 40 million downloads, reflecting its widespread adoption. Alibaba's strategic focus on open-source AI, evidenced by Qwen's 100,000+ derivative models on Hugging Face, underscores its commitment to building a global developer ecosystem. The model's deployment on edge devices like mobile phones further enhances its accessibility, enabling applications such as real-time audio descriptions for visually impaired users. Advancements Driving Global AI Innovation These releases highlight China's accelerating AI capabilities, narrowing the gap with Western counterparts. Its ability to match or surpass larger models with fewer resources demonstrates a leap in algorithmic sophistication, potentially redefining cost-performance paradigms in AI development. These advancements signal China's shift from replicating Western models to pioneering innovative architectures. Additionally, the open-source strategy of both Tencent and Alibaba amplifies their impact. By releasing Hunyuan-A13B and Qwen-VLo under permissive licenses, they empower global developers to build applications ranging from intelligent assistants to creative tools, fostering a vibrant ecosystem. This approach contrasts with the proprietary models of some Western firms, potentially accelerating China's influence in the global AI market. Alibaba's $53 billion investment in AI and cloud infrastructure over the next three years further signals its ambition to dominate the 'Model-as-a-Service' (MaaS) market, driving demand for cloud-based inference. Adversarial Dynamics in the Global AI Race It's also worth mentioning, these advancements also intensify adversarial dynamics in the U.S.-China AI race. The rapid rise of Chinese models like Hunyuan-A13B and Qwen-VLo, following DeepSeek's disruptive R1 model, has sparked concerns about the U.S.'s lead in AI innovation. DeepSeek's cost-effective training—$5.6 million compared to Meta's $60 billion AI budget—highlights China's ability to achieve competitive performance with fewer resources. This efficiency, coupled with open-source availability, challenges Western firms reliant on proprietary models and massive compute investments. However the debate of open-source remains, because the nature of these models also raises concerns about potential misuse, as unrestricted access could enable adversaries to develop advanced AI applications. Why Business Executives Should Care For business executives, these developments are a clarion call to adapt to a rapidly evolving AI landscape. Hunyuan-A13B's efficiency and Qwen-VLo's creative capabilities offer cost-effective solutions for enterprises seeking to integrate AI into operations, from automating customer service to enhancing content creation. The open-source availability reduces barriers to entry, enabling businesses of all sizes to leverage state-of-the-art AI without prohibitive costs. Western companies must innovate rapidly to maintain market share, while also navigating geopolitical risks, such as export controls and data privacy concerns. Alibaba's partnership with Apple to integrate Qwen into iPhones sold in China illustrates how Chinese AI is penetrating global markets, creating new opportunities and threats. Executives must also consider the ethical implications of open-source AI, balancing innovation with responsible use to mitigate risks like data breaches or biased outputs. The global AI market is projected to reach $1.8 trillion by 2030, and China's aggressive advancements position it as a formidable player. Businesses that fail to adopt these technologies risk falling behind competitors leveraging AI for efficiency, personalization, and innovation. Conversely, those that integrate Hunyuan-A13B and Qwen-VLo into their strategies can unlock new revenue streams, enhance customer experiences, and streamline operations. Conclusion Tencent's Hunyuan-A13B and Alibaba's Qwen-VLo represent a pivotal moment in the global AI race, blending innovation with efficiency and accessibility. While these models advance the frontier of AI capabilities, they also intensify competition and geopolitical tensions. Business executives must act swiftly to harness these technologies, aligning innovation with strategic foresight to thrive in an AI-driven world. As China's AI ecosystem continues to evolve, staying ahead requires embracing these advancements while navigating their challenges.
Mint
30-06-2025
- Business
- Mint
Alibaba unveils upgraded Qwen VLo model with image generation capabilities: How it works
Alibaba Group has rolled out an upgraded version of its artificial intelligence system, Qwen VLo, strengthening its position in the competitive landscape of AI-driven visual tools. The new model, part of the company's Qwen product line, is designed to generate and edit images using text or visual prompts. Announced in a company blog post, the enhanced Qwen VLo follows the earlier Qwen2.5-VL and introduces expanded multimodal abilities. The model now supports text-to-image and image-to-image creation, allowing users to either describe visuals or modify uploaded images using written instructions. One of its key features is 'progressive generation,' which enables users to view the step-by-step rendering of an image. For example, users can input a command like 'Generate a picture of a cute cat' or upload an image and prompt edits such as 'Add a cap on the cat's head,' with the model handling the visual modifications accordingly. The move is part of Alibaba's broader pivot towards artificial intelligence. The Hangzhou-based firm, better known globally for its online retail business, has been investing heavily in AI development through its Qwen brand. In February, the company's CEO, Eddie Wu, identified artificial general intelligence, AI with human-like reasoning, as a central goal. With Qwen VLo, Alibaba is positioning itself against both international and domestic players in the AI sector. It faces increasing competition from Chinese rivals such as DeepSeek, which recently made headlines by releasing a powerful model reportedly developed with minimal financial outlay. In response to these developments, Alibaba has continued to expand the Qwen suite to include models capable of processing text, images, audio, and video, with an emphasis on efficiency and compatibility with mobile devices. Earlier this year, the company also integrated its AI systems into an updated version of its Quark assistant app. China's AI sector is heating up as companies aggressively compete for market share. Cloud service providers, including Alibaba and Tencent, have recently reduced their prices to attract users, with DeepSeek emerging as a significant factor in this intensifying price war. Several AI startups in China have also secured substantial funding at unicorn-level valuations, adding to the competitive landscape.
