Jinse Finance reported that the Alibaba Qwen team officially released their latest research achievement - the QwQ-32B large language model. QwQ-32B, with only about 1/21 of the parameter amount of DeepSeek-R1, achieved a performance leap through reinforcement learning.

Golden Ten Data reported on April 29 that Tongyi K Man launched the kning parameter model Qwen1.5-110B for the first time, showing excellent performance in longest benchmark evaluations such as MMLU, TheoremQA, and GPQA. At present, the Qwen 1.5 series has accumulated 10 open source large models, and the number of downloads of Tongyi K Open Source models has exceeded 7 million.

According to a report by 36Kr on January 25, Alibaba Cloud announced the research progress of multimodal large models. Tongyi Qianwen visual understanding model Qwen-VL has been upgraded again, following the Plus version, the Max version has stronger visual reasoning ability and Chinese comprehension ability, can recognize people according to pictures, answer questions, create, write code, and obtain good results in multiple authoritative evaluations.

According to 36Kr reported on January 25, Alibaba Cloud announced the research progress of Long modal large models. The Qwen-VL visual understanding model has been upgraded again, following the Plus version, the Max version has been launched again, and the upgraded model has stronger visual reasoning ability and Chinese comprehension ability, can recognize people according to pictures, answer questions, create, write code, and obtain good results in Long authoritative assessments.

According to 36Kr reported on January 25, Alibaba Cloud announced the research progress of Long modal large models. The Qwen-VL visual understanding model has been upgraded again, following the Plus version, the Max version has been launched again, and the upgraded model has stronger visual reasoning ability and Chinese comprehension ability, can recognize people according to pictures, answer questions, create, write code, and obtain good results in Long authoritative assessments.

According to a report by 36Kr on January 25, Alibaba Cloud announced the research progress of multimodal large models. Tongyi Qianwen visual understanding model Qwen-VL has been upgraded again, following the Plus version, the Max version has stronger visual reasoning ability and Chinese comprehension ability, can recognize people according to pictures, answer questions, create, write code, and obtain good results in multiple authoritative evaluations.

According to Pinwan, the Arxiv page shows that Alibaba recently released an audio language model called Qwen-Audio. The model is designed to achieve universal audio comprehension by expanding the audio language pre-training to cover more than 30 tasks and various audio types, such as human voices, nature sounds, music, and songs. Research has shown that Qwen-Audio achieves significant performance on a wide range of benchmark tasks without the need for task-specific fine-tuning.

According to TechWeb's report on September 19, the domestic authoritative evaluation system Flag_ (Libra) announced the evaluation results of the latest large models on the September list. Based on the latest CLCC v2.0 subjective evaluation data set, Flag_ (Libra) September list focuses on evaluating 7 open source dialogue models that have become popular recently. Judging from the overall results, Baichuan2-13 b-chat, Qwen-7 b-chat, and Baichuan2-7 b-chat are among the best, with accuracy rates exceeding 65%. In the base model list, the objective evaluation results of Baichuan 2, Qwen, InternLM, and Aquila all surpassed the Llama and Llama2 models of the same parameter level. In the SFT model list, Baichuan 2-13 B-chat, YuLan-Chat-2-13 B, and AquilaChat-7 B rank in the top three. In both objective evaluation lists, Baichuan 2 showed excellent performance, and the basic model test surpassed Llama 2 in both Chinese and English fields. It is reported that Flag_ (Libra) is a large model evaluation system and open platform launched by Beijing Zhiyuan Artificial Intelligence Research Institute. It aims to establish scientific, fair and open evaluation benchmarks, methods and toolsets to assist researchers in comprehensively evaluating basic models and Performance of training algorithms. Flag_ The large language model evaluation system currently includes 6 major evaluation tasks, nearly 30 evaluation data sets, and more than 100,000 evaluation questions.

According to the "Kechuangban Daily", Alibaba Cloud launched a large-scale visual language model Qwen-VL today, and directly open sourced it. Qwen-VL is developed based on the 7 billion parameter model Qwen-7B of Tongyi Qianwen, which supports graphic and text input. Compared with the previous VL model, Qwen-VL not only has basic image-text recognition, description, question-and-answer and dialogue capabilities, but also adds new capabilities such as visual positioning and text understanding in images, which can be used for knowledge question-answering and image caption generation , image question and answer, document question and answer, fine-grained visual positioning and other scenarios.

Jinse Finance reported that technology giant Alibaba Group announced on August 3 that its cloud computing division released two open-source artificial intelligence (AI) models. Its two large language models (LLMs) are called Qwen-7B and Qwen-7B-Chat, each with 7 billion parameters. The new model is designed to help bring artificial intelligence into the operations of small and medium-sized businesses. Qwen-7B and Qwen-7B-Chat have various features that are attractive to enterprises, such as enabling "free access to code, model weights, and documentation by academics, researchers, and commercial organizations around the world," the company said. On Aug. 1, the company also announced a vector engine update to its AnalyticDB data warehouse service, which will allow its enterprise customers to quickly create custom generative AI applications.

According to 36 Krypton reports, on August 3, Alibaba Cloud open-sourced Tongyi Qianwen’s 7 billion parameter model, including the general model Qwen-7 B and the dialogue model Qwen-7 B-Chat. , free, and commercially available. This move makes Alibaba Cloud the first large-scale technology company in China to join the ranks of large-scale open source.