ByteDance Open Sources MagicAnimate: A New Tool for AI-Generated Dance Videos

ByteDance Launches Top Seed Program to Recruit AI Talent from the Class of 2026

ByteDance recently announced the official launch of its "Top Seed" program for recruiting top AI talent from the class of 2026. The program aims to recruit approximately 30 outstanding doctoral students. This initiative focuses on cutting-edge artificial intelligence, encompassing research areas such as large language models, machine learning algorithms and systems, multi-modal generation and understanding, and speech processing. ByteDance hopes to attract young talents with strong potential and passion in the field of large language model research. Unlike previous recruitment plans, this year's "Top Seed" program emphasizes no restrictions on academic background.

ByteDance Unveils QuaDMix: A Unified Framework for Large Language Model Pre-training Data Quality and Diversity

ByteDance Launches Vidi, a Multimodal Model Leading the Trend in Ultra-Long Video Understanding and Editing

ByteDance Releases Efficient Pre-training Length Scaling Technology, Breaking Through Long Sequence Training Bottlenecks

ByteDance Restructures AI Product Line: Cat Box Leadership Change, Xinghui Merged into Doubao, Focusing on Growth

According to LatePost, ByteDance recently made significant adjustments to its AI product department, Flow. The social companionship AI product, Cat Box, has a new leader. The previous head, Liang Chenqi, has left the company, and has been replaced by Xi Yuan (codename), the former head of Xinghui. Meanwhile, the Xinghui team, which develops AI camera and image generation applications, is slated to merge into the Doubao App, under the unified management of Doubao App's head, Lu You (codename). The Flow department is headed by Zhu Jun and includes Doubao, Cat Box, Xinghui, Doubao Aixue, and G.

ByteDance Research Open-Sources ChatTS-14B: Native Understanding and Reasoning Over Time

ByteDance Research has announced the open-sourcing of ChatTS-14B, a 14-billion parameter large language model (LLM) specifically designed for understanding and reasoning with time series data. Released under the Apache2.0 license, ChatTS-14B's open-source release has garnered significant attention within the AI community, marking a substantial advancement in the intersection of time series analysis and generative AI. ChatTS-14B: An Intelligent Conversational Engine for Time Series. ChatTS-14B is based on Qwen2.5-1...

Coze Space Officially Opens Beta Testing, Supporting MCP Extension Integration

ByteDance's technology team announced that its new AI collaborative workspace, "Coze Space", is officially opening beta testing. Coze Space aims to be the optimal place for users to collaborate with AI Agents, providing comprehensive services ranging from answering questions to solving problems, helping users work more efficiently.

BMW Brilliance and ByteDance's Volcano Engine Partner to Drive AI-Powered Automotive Marketing

Recently, BMW Brilliance Lynk & Co Digital Information Technology Co., Ltd. (Lynk & Co) and ByteDance's Volcano Engine have partnered to innovate automotive marketing services with the help of Artificial Intelligence (AI) technology. This collaboration leverages AI to achieve precise product matching and purchase recommendations, optimize content guidance, and enhance the user car-buying experience and dealer operational efficiency. BMW Group President and CEO in Greater China, Gao Xiang, stated that AI is key to BMW's creation of smarter and more considerate mobility solutions, and is being rapidly integrated into R&D, production, supply chain, product, service, and operations.

ByteDance Releases UI-TARS-1.5: Open-Source Multimodal Agent Leading a New Wave in GUI Automation

ByteDance has officially released UI-TARS-1.5 on the Hugging Face platform, an open-source multimodal agent built upon a powerful vision-language model. This release marks another significant breakthrough for ByteDance in the field of AI automated interaction, providing developers and users with a highly efficient and intelligent cross-platform GUI (Graphical User Interface) automation solution. UI-TARS-1.5: A New Benchmark for Multimodal Agents. UI-TARS-1.5 is the latest in ByteDance's UI-TARS series...

ByteDance Doubao Open-Source Seed Agent Model UI-TARS-1.5

The ByteDance Doubao large model team announced the open-sourcing of UI-TARS-1.5, an open-source multimodal agent built on a vision-language model capable of efficiently executing various tasks in a virtual world. The model achieved state-of-the-art (SOTA) performance on seven typical GUI (Graphical User Interface) benchmark evaluations and demonstrated, for the first time, its long-term reasoning capabilities in games and interactive capabilities in open spaces. This open-source project marks a significant advancement in multimodal agent technology for GUIs.