Skip to content

TeleChat to Xing4: Advancing Frontier Agentic Models on Domestic Chips and Frameworks

Xingchen Agi Lab Baotian Hu Bing-Kai Yang Bo-Yang Wang Chang-Zai Pan Chao Lyu Chao Wang Chen Li Chen-Shuo Pan Chen-Yu Lin Chen-Yuan Zhang Chong-Ming Liu Da-Wei Fan Di Wu Dong Guo Dong Li Dongfang Li En-Lin Liu Fan-Ye Xiao Feng-Kai Li Fu-Bei Yao Gang Lin Han Gao Han-Ming Wu Hao-Wei He Hua-Peng Liu Hua Wang Hui-Nan Xu Jia-Hong Xiong Jia-Xin Peng Jian-Xiang Ma Jian-Xing Zhuge Jie Zhang Ji Xiong Jing-Lin Chen Jing-Qi Li Jing-Xuan Li Ji-Wang Ye Jun-Chen Li Jun-Biao Li Jun-Yang Wu Kai-Dong Yu Kai Jia Ke-Hai Chen Kuai Li Lao-Bin Liu Lei Li Li-Bo Qin Ling-Ling Shi Lu-Wen Pu Meng-Hao Xu Meng-Jiao Bao Meng-Qi Chen Min Zhang Ming-Rui Wang Na-Ying Wang Pei-Cong Han Peiming Guo Peng-Fei Li Qiang Feng Qi Wang Qing-Jun Zhu Qing-Li Li Qi-Tai Ding Qi-Yi Xie Quan-Chao Zhao Rong-Bao Han Rui Xue Rui-Yu Fang Rui Song Shi-Jie Li Shi-Quan Wang Shuang-Yong Song Si-Shi Xiong Teng Su Wan-Kang Bao Wei-Jie Zhang Wei Mao Wei-Wei Fu Wei-Yang Huang Wen-Ming Deng Wen-Qing Hu Wu-Jie Zhong Xi Yu Xiao-Hui Hu Xiao-Min Fang Xiao-Meng Huang Xiao-Qiong Shi Xiao-Tong Shang Xiao-Yu Li Xing-Lei Xu Xin Jiang Xin Wang Xin Zhang Xin-Zhang Liu Xue Liu Xue-Feng Bai Xue-Wei He Ya-Han Li Yan-Han Zhang Yan-Xue Xue Yan-Zhang Wu Yang Yu Yi Liu Yin Zhang Ying-Yan Li Yi-Tong Yao Yi-Xuan Li Yong-Qiang Xia Yong-Xiang Li Yin-Zhou Zhu Yu Zhao Yu-Hao Jiang Yun-Peng Zhang Yu-Ting Lin Yuxuan Wu Yu-Yao Huang Yu-Yi Wang Ze-Bo Peng Zhen-Wu Pei Zhen-Xiong Li Zhao Wang Zhao-Xi Zhang Zhi-Hao Yang Zhi-Qiang Dong Zhi-Wei Yan Zhi-Long Lu Zhong-Hao Guo Zhong-Heng Wu Zhong-Jiang He Zhong-Qiu Li Zhu Yuan Zhuo-Ru Jiang Zhuo-Yi Wei Zi-Jia Guo Zi-Lu Huang Zhi-Peng Guo Zi-Rui Li Zi-Han Wang
Nov 2026 · 0 citations · 30 references
Computer Science

TL;DR

This report presents Xing 4, a series of large language models that extends the TeleChat lineage through coordinated advances in model architecture and training infrastructure on Ascend, and introduces TeleGym, a unified agent execution platform that connects execution, evaluation, and trajectory construction to policy optimization.

Abstract

We present Xing 4, a series of large language models that extends the TeleChat lineage through coordinated advances in model architecture and training infrastructure on Ascend. This report describes the design principles and systems engineering behind the series, spanning architecture selection, large-scale pretraining, and reasoning and agentic post-training. We study how sparse activation, attention design, model depth and width, and speculative decoding balance model capacity with computation, memory access, and execution overhead. Our pretraining infrastructure builds on MindSpore HyperParallel, combining topology-aware hybrid sharding, efficient distributed optimizer execution, and coordinated management of long-sequence workloads and activation memory. CPU affinity and core isolation further reduce host scheduling interference. Together, these mechanisms support the pretraining of a 285B-parameter model on 8,192 Ascend 910C cards, providing a concrete setting for the architectural and systems studies. For post-training, we introduce TeleGym, a unified agent execution platform that connects execution, evaluation, and trajectory construction to policy optimization. Disaggregated asynchronous execution, bounded oversampling, and partial rollout accommodate heterogeneous task durations while preserving trajectory continuity across policy updates. Together, these advances establish an Ascend-oriented training foundation for the continued development of Xing 4 across model scales and increasingly demanding reasoning and agent workloads.

View source

Similar papers

#artificial intelligence Preprint Jul 2024

The Llama 3 Herd of Models

It is found that Llama 3 delivers comparable quality to leading language models such as GPT-4 on a plethora of tasks, and performs competitively with the state-of-the-art on image, video, and speech recognition tasks.

Aaron Grattafiori, Abhimanyu Dubey, Abhinav Jauhri et al. · 21,082 citations · ⚡3,832
#artificial intelligence Conference Open access Jun 2018

Gibson Env: Real-World Perception for Embodied Agents

This paper investigates developing real-world perception for active agents, proposes Gibson Environment for this purpose, and showcases a set of perceptual tasks learned therein.

F. Xia, Amir Zamir, Zhi-Yang He et al. · 1,069 citations · ⚡112
#artificial intelligence Open access May 2023

Evaluating the Performance of Large Language Models on GAOKAO Benchmark

GAOKAO-Bench is introduced, an intuitive benchmark that employs questions from the Chinese GAOKAO examination as test samples, including both subjective and objective questions that contribute a robust evaluation benchmark for future large language models and offers valuable insights into the advantages and limitations...

Xiaotian Zhang, Chun-yan Li, Yi Zong et al. · 216 citations · ⚡17

KTO: Model Alignment as Prospect Theoretic Optimization

This work proposes a HALO that directly maximizes the utility of generations instead of maximizing the log-likelihood of preferences, and suggests that the best loss depends on the inductive biases most appropriate for a given setting, an oft-overlooked consideration.

Kawin Ethayarajh, Winnie Xu, Niklas Muennighoff et al. · 195 citations · ⚡28
#artificial intelligence Open access Jul 2024

Gender, Race, and Intersectional Bias in Resume Screening via Language Model Retrieval

This work investigates the possibilities of using LLMs in a resume screening setting via a document retrieval framework that simulates job candidate selection and finds that the MTEs are biased, significantly favoring White-associated names in 85% of cases and female-associated names in only 11.1% of cases.

Kyra Wilson, Aylin Caliskan · 131 citations · ⚡8

Related blog posts

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.