Alibaba Cloud has open-sourced over 100 new AI models, collectively known as Qwen 2.5, as announced during the company’s annual Apsara Conference. This move highlights Alibaba Cloud’s commitment to driving innovation and expanding access to AI technologies for businesses and developers worldwide.
In addition to the AI models, Alibaba Cloud has introduced a full-stack infrastructure overhaul to meet the growing demand for powerful AI computing. This revamped infrastructure includes cutting-edge cloud products and services that enhance computing, networking, and data center architecture, all designed to support the development and application of AI models on a global scale.
Eddie Wu, Chairman and CEO of Alibaba Cloud Intelligence, said: “Alibaba Cloud is investing, with unprecedented intensity, in the research and development of AI technology and the building of its global infrastructure. We aim to establish an AI infrastructure of the future to serve our global customers and unlock their business potential.”
The newly-released Qwen 2.5 models range in size from 0.5 to 72 billion parameters, with improved capabilities in areas such as mathematics and coding. These models support over 29 languages, making them suitable for diverse AI applications across sectors like automotive, gaming, and scientific research.
Related: Microsoft & BlackRock Partnered for a $30 Billion AI Infrastructure Fund
Alibaba Cloud’s Open-Source AI Models Gain Momentum
Since its launch in April 2023, the Qwen model series has seen tremendous success, with over 40 million downloads on platforms like Hugging Face and ModelScope. Additionally, over 50,000 derivative models have been created on Hugging Face, highlighting the growing influence of Alibaba Cloud’s AI offerings.
Jingren Zhou, CTO of Alibaba Cloud Intelligence, commented: “This initiative is set to empower developers and corporations of all sizes, enhancing their ability to leverage AI technologies and further stimulating the growth of the open-source community.”
Along with open-source models, Alibaba Cloud also announced upgrades to its proprietary Qwen-Max model, which now offers improved performance in areas like language comprehension, reasoning, math, and coding, bringing it on par with other state-of-the-art AI models.
Related: What is Shadow AI?
Expanding Multimodal Capabilities
Alibaba Cloud has also strengthened its multimodal AI capabilities by introducing a new text-to-video model as part of its Tongyi Wanxiang large model family. This model can generate high-quality videos in various styles, from realistic scenes to 3D animations, based on text inputs in both Chinese and English.
Additionally, the newly launched Qwen2-VL vision-language model supports video-based question answering and can analyze videos lasting over 20 minutes, showcasing Alibaba Cloud’s continued push to expand the limits of AI technology.
AI Developer Assistant for Programmers
Alibaba Cloud also launched AI Developer, an AI assistant powered by Qwen, designed to help programmers automate key tasks such as requirement analysis, code programming, and bug identification and fixing. This tool is set to streamline development processes and enhance productivity across the board.
Supporting AI with Next-Gen Infrastructure
To support these advancements in AI, Alibaba Cloud has made several key infrastructure upgrades:
- CUBE DC 5.0: A next-gen data center architecture that boosts energy and operational efficiency.
- Alibaba Cloud Open Lake: A solution designed to maximize data utility for generative AI applications.
- PAI AI Scheduler: A proprietary cloud-native scheduling engine for better management of computing resources.
- DMS: OneMeta+OneOps: A platform offering unified metadata management across multi-cloud environments.
- 9th Generation Enterprise Elastic Compute Service (ECS): This new instance offers improved performance for various applications, helping customers maximize the potential of AI.
With these upgrades and the release of over 100 open-source models, Alibaba Cloud aims to equip its customers and partners with the tools they need to build more efficient, sustainable, and inclusive AI applications.
For more tech, security, and marketing insights, visit Daily Digital Grind. If you’re interested in contributing, check out our Write for Us page to submit your guest posts!