About Qwen
Alibaba Cloud provides Tongyi Qianwen (Qwen) model series to the open-source community. This series includes Qwen, the large language model (LLM); Qwen-VL, the large language vision model; Qwen-Audio, the large language audio model; Qwen-Coder, the coding model; and Qwen-Math, the mathematical model. You can try Qwen models and easily customize and deploy them in Alibaba Cloud Model Studio.
The latest Qwen 2.5 models are pre-trained on our latest large-scale dataset, which includes up to 18 trillion tokens. Compared to Qwen2, Qwen2.5 has acquired significantly more knowledge (MMLU: 85+) and has greatly improved capabilities in coding (HumanEval 85+) and mathematics (MATH 80+). Additionally, the new models have significantly improved in following instructions, generating long texts, understanding structured data, and generating structured outputs. Qwen2.5 models are generally more resilient to the diversity of system prompts, enhancing role-play implementation and condition-setting for chatbots. Qwen2.5-Coder has been trained on 5.5 trillion tokens of code-related data, delivering competitive performance against larger language models on coding evaluation benchmarks. Qwen2.5-Math supports both Chinese and English and incorporates various reasoning methods, including Chain-of-Thought (CoT), Program-of-Thought (PoT), and Tool-Integrated Reasoning (TIR).
-
Leading Performance in Multiple Dimensions
Qwen outperforms other open-source baseline models of similar sizes on a series of benchmark datasets that evaluate natural language understanding, mathematical problem-solving, coding, etc.
-
Easy and Low-Cost Customization
You can deploy Qwen models with a few clicks in PAI-EAS, and fine-tune them with your data stored on Alibaba Cloud, or external sources, to perform industry or enterprise-specific tasks.
-
Applications for Generative AI Era
You can leverage Qwen APIs to build generative AI applications for a broad range of scenarios such as writing, image generation, audio analysis, etc. to improve work efficiency in your organization and transform customer experience.
Try Qwen Models on Alibaba Cloud Model Studio
Customer Success Stories
"Working closely with Alibaba Cloud, we managed to harness the benefits of the Qwen LLM and Dedicated Model Studio and vastly improved the efficiency of generating adverse event reports from huge amounts of medical literature. We’re proud that we have pioneered this innovation in the industry. We expect to explore more AI-based innovations with Alibaba Cloud."
Xin Zhong, IT head of AstraZeneca China
AstraZeneca is a global biopharmaceutical company focused on science-led innovation. Working with Alibaba Cloud, AstraZeneca's team has built the industry’s first adverse event summary system, which makes full use of the AI engineering capabilities provided by Alibaba Cloud's Dedicated Model Studio and Tongyi Qwen LLM. As the system can quickly comb through vast amounts of medical literature and formulate output documents according to the requirements, the efficiency in creating adverse event reports has increased by 300%, and the accuracy has increased from 90% to 95%.
"Regarding language capabilities, we found that Alibaba Cloud’s Tongyi Qianwen (Qwen) not only performed well in English, but also proved to be the best publicly available option for supporting Japanese. We chose Qwen because our LLM’s accuracy significantly improved when fine-tuned with a base model capable of understanding Japanese."
Shunichi Taniguchi, Director | Senior Researcher, Lightblue Co., Ltd.
Lightblue Co., Ltd. is a startup dedicated to democratizing AI. Alibaba Cloud’s Tongyi Qianwen (Qwen) was crucial for the release of Lightblue’s Karasu and Qarasu models, as they gave Lightblue the capabilities needed to succeed in Japanese language processing. Qwen’s advanced architecture and extensive training in East Asian languages provided outstanding accuracy when fine-tuned for Japanese, ensuring clear and relevant interactions.
"Selecting Alibaba Cloud Services as our partner was a wise decision rooted in their demonstrated expertise in data analysis, AI services, and language models. Their understanding of applications and services has been pivotal in enhancing our consumer experience and technology innovation."
Tina Chen, Chief Digital Officer, Shiseido China
Drunk Elephant (acquired by Shiseido) is one of the fastest-growing skincare brands in the last decade, established in the United States in 2013. To answer Drunk Elephant’s needs, Shiseido worked with Alibaba Cloud to build “Drunk GPT,” an interactive multi-round dialogue service powered by Alibaba Cloud’s Tongyi Qwen LLM. Drunk GPT understands text, audio, and images and focuses on skincare knowledge, relevant Q&A, product consultation, and product recommendations, resulting in improved online marketing and security.
"We are honored to collaborate with Alibaba Cloud in this venture. Haleon's personalized nutrition platform, the good deposit Keyijia, will be able to provide consumers with efficient and personalized health support services empowered by Haleon's proprietary Knowledge Graph and the AI nutrition assistant enabled by Qwen."
Susan Gu, General Manager of Haleon Mainland China and Hong Kong
Haleon is a global leader in consumer health, with a purpose to deliver better everyday health with humanity. By harnessing Alibaba Cloud's Qwen models, and integrating them with Haleon's proprietary knowledge graph through retrieval-augmented generation, Haleon’s solution seamlessly integrates across all customer touchpoints, ensuring a cohesive end-to-end journey and hyper-personalized healthcare management. Now, a single nutritionist can serve over 1,300 consumers, representing a six-fold increase in efficiency compared to traditional models.
Chatbot based on Qwen, Qwen-Audio, and Qwen-VL answering questions containing multimodal data
Qwen-VL producing a cartoon-style human portrait based on prompts
Qwen-VL recognizing the objects (the woman and the dog) in the image and their gestures (high five)
Qwen-Audio analyzing the identity and emotions of the speaker and recommending replies
Qwen2.5-72B providing formatted output based on the requirement and input data (table in JSON format)
Qwen2.5-72B generating JSON code step by step with explanations as requestedQwen2.5-72B writing a report of over 5,000 Chinese characters on the requested subject
Qwen-Agent: Developing AI Agents and Applications in Simple Steps
Qwen-Agent is a framework for developing LLM applications based on the instruction following, tool usage, planning, and memory capabilities of Qwen models. It provides various components for LLMs, prompts, and agents. Follow this tutorial and learn to use the Assistant component to add customized tools and quickly develop an agent that uses these tools.
Download Qwen models and try Qwen demos on Hugging Face.
Qwen on Hugging Face >
Download Qwen models and try Qwen demos on ModelScope.
Qwen on ModelScope >
Download Qwen models and read tutorials on GitHub.
Qwen on GitHub >
