Hunyuan Large Model

On July 14, according to news, the free version of Hy3 has come to the top of OpenRouter LLM week. On July 6th, the telecommunication mixer officially launched the Hy3 model and subsequently launched a two-week free API call on the AI Model Convergence platform OpenRouter. According to the data, the free version of Hy3 processed 6.13 trillion token per week, the highest on the weekly list. This figure exceeds the 5.95 trillion Token of MiMo-V2.5..

According to an April 29 report by Interface News, Tencent has carried out a comprehensive restructuring of its hybrid big model research and development system, mainly focusing on the three core sectors of arithmetic, algorithms, and data, and promoting the continuous development of the hybrid big model by optimizing the deployment of the team and increasing R&D investment. According to reports, Tencent set up two new departments after this restructuring: the Department of Large Language Models and the Department of Multimodal Models, which are responsible for exploring the cutting-edge technologies of large language models and multimodal large models respectively, continuously iterating the basic models, and improving the modeling capabilities. At the same time, it further strengthens the big model data capability and platform base construction, in which the data platform department focuses on big model data all...

According to Tencent's March 1 news, Tencent's AI assistant for consumers based on the hybrid turbo model "Tencent Yuanbao computer version" is now officially online, providing Windows and macOS versions. 1AI learned that Tencent Yuanbao is an AI product based on Tencent's hybrid turbo model, and also supports dual-mode switching DeepSeek model. The assistant mainly provides AI search, AI summarization, AI writing and other core capabilities, and is able to parse multiple WeChat public number links, URLs, and PDF,...

Tencent announced on December 3, mixed yuan large model on-line and open source literate video capabilities, the number of participants 13 billion, support for bilingual input in English and Chinese. Officials claimed that Tencent hybrid video generation model can generate "ultra-realistic" high-quality video, generated by the screen is not easy to deform; in the mirror or mirror scene, you can do the mirror reflection action and the outside completely synchronized, light and shadow reflections basically in line with the laws of physics. According to reports, Tencent hybrid video generation model using DiT architecture, adapted to a new generation of text encoder to improve semantic compliance, better response to multiple subjects depicted, to achieve more detailed finger...

Tencent's public relations director Zhang Jun announced on the evening of November 28, Tencent mixing big model December 3 officially on-line video generation capabilities, users can submit an application for trial in the Tencent Yuanbao App. Enterprise customers through the Tencent cloud to provide service access, the current API synchronization open internal test applications. At the same time, Tencent will also open source the video generation big model, the number of participants 13 billion. In the middle of this month, Tencent large model App Yuanbao released version 2.0, interface upgrades include updating the dialog list, support for the historical use of asset deposition, users can more easily find the history of...

Introduction to Hunyuan DIT Model In previous articles, the Hunyuan model open sourced by Tencent has been introduced many times. The core advantages of Tencent's Hunyuan model include: Chinese element understanding: Hunyuan DiT provides bilingual generation capabilities, and has advantages in understanding Chinese elements. Especially the understanding of Chinese and Chinese culture (for example: ancient poetry). Long text understanding ability: Hunyuan DiT can analyze and understand the information in long texts and generate corresponding works of art. Support longer and more complex prompt text understanding. Fine-grained semantic understanding: Hunyuan DiT can capture the subtleties in the text, thereby generating perfect fit for user needs...

Tencent Questionnaire announced the launch of a full-process AI solution, based on Tencent's hybrid big model, to help enterprises improve the quality and efficiency of research. Tencent questionnaire is Tencent's online questionnaire survey platform, the service covers questionnaire survey, information reporting, online assessment and other work scenes, to help users recover 3.4 billion questionnaires. Through the access of Tencent's hybrid big model, Tencent questionnaire can provide AI solutions in questionnaire design, placement, recovery and analysis. In the questionnaire design process, users can describe their needs in natural language, and AI can generate customized questionnaires and check the quality to improve the filling experience. In the placement process...

At the Tencent Cloud Generative AI Industry Application Summit, Tencent said that it will officially release a new app for the C-end - "Tencent Yuanbao" on May 30th. Behind this app is Tencent's chain-wide self-developed universal large language model, the Mixed Yuan Large Model, which demonstrates excellent Chinese language comprehension and creation capabilities, logical reasoning capabilities, and reliable task execution capabilities with its parameter scale of more than a trillion. Tencent's Mixed Meta Model reaches such heights thanks to its adoption of the Model of Mixed Experts (MoE) structure. This innovation not only dramatically improves the performance of the model, but also significantly reduces the cost of reasoning...

Tencent Holdings today announced its financial results for the fourth quarter and full year of 2023 ended December 31, with various figures showing different growth trends. In the fourth quarter, Tencent Holdings' revenue reached RMB155.2 billion, a year-over-year increase of 7%. Looking at the figures for each business segment in the fourth quarter, value-added services revenue amounted to RMB69.1 billion, a slight decline of 2% compared to the same period a year ago. among the value-added services, game revenue from the international market amounted to RMB13.9 billion, realizing year-on-year growth of 1%; revenue from the local games amounted to RMB27.0 billion, a year-on-year decline of 3%; and revenue from social networks amounted to RMB1.3 billion. 27 billion, down 3% year-on-year; social networking...

At the plenary meeting of the Artificial Intelligence Subcommittee of the National Information Technology Standardization Technical Committee on December 22, Tencent Hunyuan Big Model was the first to pass the national big model standard compliance test. This evaluation was initiated by the China Electronics Technology Standardization Institute and tested the main domestic big models based on 32 subdivision dimensions. Determination: Tencent Hunyuan Big Model meets the relevant technical requirements of the language big model in "Artificial Intelligence Large Scale and Training Model Part 2: Evaluation Indicators and Methods". In addition, Tencent Cloud also initiated the discussion and compilation of the "Artificial Intelligence Model as a Service (MaaS) Functional Requirements" as the lead proposal...

Tencent Hybrid Big Model is a powerful big language model developed by Tencent. It has a number of core advantageous capabilities, including multi-round dialog, content creation, logical reasoning and knowledge enhancement, with extensions for multi-modal capabilities. These selling points will be introduced one by one below. The Tencent Hybrid Grand Model has a wide range of application scenarios. In document scenarios, it can provide document creation, text touch-up, text proofreading, and formulas and charts generation capabilities to improve creation efficiency and experience. In meeting scenarios, it can provide in-meeting Q&A, meeting summarization, meeting to-do list organization and other capabilities to simplify meeting operations and improve meeting efficiency. In advertising scenarios...