Tongyi AI, a subsidiary of Alibaba, publicly clarified recent marketing investment data for AI products that was circulated online. Previously, it was reported that multiple AI products had advertising expenditures reaching the hundred million level in the third quarter, which attracted significant attention, particularly regarding specific figures related to Tongyi Qianwen. According to Tongyi AI's official response, the rumored figures of 4.5 million investment for Tongyi Qianwen in Q3 and 900,000 in the first 20 days of October are severely inconsistent with the actual situation. The official statement clearly indicated that the actual investment is far lower than the rumored figures, but did not disclose the real investment amount.

During the opening ceremony of the 2024 Yunqi Conference held today, Wu Yongming, director and CEO of Alibaba Group, as well as chairman and CEO of Alibaba Cloud Intelligent Group, delivered a speech sharing profound insights on the development of Artificial Intelligence (AI). He revealed that over the past year, the call price for the Tongyi Qianwen API on Alibaba Cloud's Bai Lian has dramatically dropped by 97%, with the cost of calling one million tokens now as low as 0.5 yuan, and Alibaba Cloud will continue to strive to reduce costs.

The Tongyi Qianwen team announced the official open source of the latest member of the Qwen family, the Qwen2.5 series language models, three months after the Qwen2 release. This marks one of the largest open source releases in history, including the general language model Qwen2.5, as well as the Qwen2.5-Coder and Qwen2.5-Math models specifically designed for programming and mathematics.

On August 30, 2024, Alibaba Damo Academy's Tongyi Qianwen team announced a significant update to its latest achievement—the Qwen2-VL model. The Qwen2-VL model has achieved remarkable improvements in image understanding, video processing, and multilingual support, setting new benchmarks for key performance indicators.

Alibaba's Tongyi Qianwen team has released a math model named Qwen2 Math Demo, which excels at processing mathematical problems in text input and formulas in images and screenshots. It even surpasses GPT-4 on the MATH dataset, improving the score by 9.6%. The model has three versions, with the 72B version being particularly outstanding, outshining GPT-4 in challenging math problems and also outperforming Llama-3.1 on classic test sets like GSM8K and MATH.

Tongyi Qianwen has launched a new domain 'tongyi.ai' and introduced new features including deep search capabilities to enhance user experience. The new web-based chat function supports deep search, providing professional and structured search results, displaying source webpages through digital badges, and offering one-click jump links. The app upgrade also adds an image micro motion effect feature that supports multi-size image uploads, optimizes the custom performance function, solving long audio upload issues by extending upload duration from 120 seconds to 300 seconds, while also introducing 3:4 aspect ratio support to meet user needs.

In the early hours of today, Alibaba's Tongyi Qianwen team released the Qwen2 series of open-source models. This series includes pre-trained and instruction-tuned models in five sizes: Qwen2-0.5B, Qwen2-1.5B, Qwen2-7B, Qwen2-57B-A14B, and Qwen2-72B. Key information indicates that these models show significant improvements in parameter count and performance compared to the previous generation Qwen1.5.

Recently, the Aliyun Tongyi Qianwen APP launched a free feature called 'Tongyi Dance King'. Users only need to input a specified password and upload a photo to generate a personalized dance video. The system retains original image features, including facial expressions, body proportions, clothing, and more. It provides 12 popular dance templates, such as Subject Three, Mongolian Dance, and Paddle Dance. This is powered by the self-developed video generation model 'Animate Anyone' from the Aliyun Tongyi Laboratory.

Alibaba Cloud Tongyi Qianwen 2.0 has been officially released, reaching a parameter scale of hundreds of billions. Alibaba Cloud's Chief Technology Officer, Zhou Jingren, announced a comprehensive upgrade of the Alibaba large model family. The Tongyi Qianwen 2.0 model achieves significant improvements and supports features such as voice conversation. Users can download and experience Alibaba Cloud Tongyi Qianwen 2.0, opening a new chapter in artificial intelligence.

The Alibaba Cloud Tongyi Qianwen large model has been officially opened, allowing users to experience the Alibaba Cloud Tongyi Qianwen API through the official website. It is also open to enterprise users, with over 200,000 enterprise and organization users applying for access. Partners such as OPPO, Dewu, and DingTalk have already collaborated with Alibaba Cloud. Alibaba Cloud will also open source a larger parameter version of the large model.

Following the open sourcing of the general model Qwen-7B in August, Alibaba Cloud has released the visual language model Qwen-VL. Qwen-VL supports both Chinese and English and can perform tasks such as image captioning, image Q&A, and document Q&A. Qwen-VL also enables open-domain visual localization in Chinese by marking detection boxes within images.

Qwen-VL is a large-scale vision language model launched by Aliyun that supports image and text input. Compared to other vision language models, Qwen-VL has added capabilities such as visual localization and image text understanding. Qwen-VL has received over 3,400 stars on GitHub and has been downloaded more than 400,000 times.

{title: Alibaba releases fiscal report for Q1 of 2024, with quarterly revenue of 234.16 billion yuan, Alibaba Cloud revenue at 25.123 billion yuan, and Tongyi Qianwen's cumulative users exceeding 550,000. list: Alibaba's quarterly revenue is 234.16 billion yuan, a year-on-year increase of 14%. Alibaba Cloud's revenue is 25.123 billion yuan, a year-on-year increase of 4%. The total number of users on Tongyi Qianwen's platforms has exceeded 550,000.}