Current status, applications, and prospects of 6G wireless network intelligence driven by large models and agents

doi:10.12267/j.issn.2096-5931.2026.10.009

Abstract

Abstract:

To address the increasingly complex service requirements and operating environments of 6G wireless networks, conventional task-specific models face limitations in cross-task adaptation and domain knowledge integration. Large models, with their capabilities in knowledge representation and task generalization, offer a promising paradigm for advancing wireless network intelligence. This paper reviews recent advances in foundation model development for wireless networks, wireless task applications, network agents, and capability evaluation. Furthermore, through a case study on user experience diagnosis in wireless networks, we demonstrate an application paradigm in which large models collaborate with domain knowledge and specialized tools to complete complex tasks. Finally, this paper discusses key challenges, including trustworthy closed-loop operation, cross-scenario transfer, evaluation frameworks, and the industrial ecosystem.

Key words: wireless network intelligence, 6G, large model, network agent

CLC Number:

TN929.11

LI Qizhe, SUN Qianyi, LIU Chenji, YANG Wu, ZHU Guangxu. Current status, applications, and prospects of 6G wireless network intelligence driven by large models and agents[J]. Information and Communications Technology and Policy, 2026, 52(10): 72-79.

Add to citation manager EndNote|Ris|BibTeX

URL:

http://ictp.caict.ac.cn/EN/10.12267/j.issn.2096-5931.2026.10.009

http://ictp.caict.ac.cn/EN/Y2026/V52/I10/72

Figures/Tables 2

References 32

[1]	NGMN Alliance. ITU-R framework for IMT-2030: Review and future direction[R]. 2024-02-14.
[2]	TS 23.288 V18.11.0 Architecture enhancements for 5G System (5GS) to support network data analytics services[S].
[3]	Zhang C, Patras P, Haddadi H. Deep learning in mobile and wireless networking: a survey[J]. IEEE Communications surveys & tutorials, 2019, 21(3): 2224-2287.
[4]	Boutaba R, Salahuddin M A, Limam N, et al. A comprehensive survey on machine learning for networking: evolution, applications and research opportunities[J]. Journal of Internet Services and Applications, 2018, 9:16. doi: 10.1186/s13174-018-0087-2
[5]	Ayoubi S, Limam N, Salahuddin M A, et al. Machine learning for cognitive network management[J]. IEEE Communications Magazine, 2018, 56(1):158-165.
[6]	Wang C X, Renzo M D, Stanczak S, et al. Artificial intelligence enabled wireless networking for 5G and beyond: recent advances and future challenges[J]. IEEE Wireless Communications, 2020, 27(1):16-23.
[7]	Chang Y, Wang X, Wang J, et al. A survey on evaluation of large language models[J]. ACM transactions on intelligent systems and technology, 2024, 15(3):1-45.
[8]	Boateng G O, Sami H, Alagha A, et al. A survey on large language models for communication, network, and service management: application insights, challenges, and future directions[J]. IEEE Communications Surveys & Tutorials, 2025, 28:527-566.
[9]	Zou H, Zhao Q, Tian Y, et al. TelecomGPT: a framework to build telecom-specific large language models[J]. IEEE Transactions on Machine Learning in Communications and Networking, 2025, 3:948-975. doi: 10.1109/TMLCN.2025.3593184 URL
[10]	Lee S, Arya D, Cho S M, et al. Telbench: a benchmark for evaluating telco-specific large language models[C]// Proceedings of the 2024 conference on empirical methods in natural language processing:Industry Track. Miami: Association for Computational Linguistics, 2024: 609-626.
[11]	Wu D, Wang X, Qiao Y, et al. Netllm: adapting large language models for networking[PP/OL]. arXiv(2024-08-06)[2026-09-01]. https://arxiv.org/abs/2402.02338.
[12]	Shao J, Tong J, Wu Q, et al. WirelessLLM: empowering large language models towards wireless intelligence[PP/OL]. arXiv(2024-06-15)[2026-09-01]. https://arxiv.org/abs/2405.17053.
[13]	Khan M Z, Ge Y, Ullah U, et al. TelcoGPT: a hybrid embedding approach for telecom-specific Q&A and code retrieval[C]// IEEE INFOCOM 2025-IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS). London: IEEE, 2025: 1-6.
[14]	Alikhani S, Charan G, Alkhateeb A. Large wireless model (LWM): a foundation model for wireless channels[PP/OL]. arXiv(2025-04-07)[2026-09-01]. https://arxiv.org/abs/2411.08872.
[15]	Liu B, Gao S, Liu X, et al. WiFo: Wireless foundation model for channel prediction[J]. Science China Information Sciences, 2025, 68(6): 162302. doi: 10.1007/s11432-025-4349-0
[16]	Panitsas I, Tassiulas L. A family of open time-series foundation models for the radio access network[PP/OL]. arXiv(2026-04-05)[2026-09-01]. https://arxiv.org/abs/2604.04271.
[17]	Qiu K, Bakirtzis S, Wassell I, et al. Large language model-based wireless network design[J]. IEEE Wireless Communications Letters, 2024, 13(12): 3340-3344. doi: 10.1109/LWC.2024.3462556 URL
[18]	Noh H, Shim B, Yang H J. Adaptive resource allocation optimization using large language models in dynamic wireless environments[J]. IEEE Transactions on Vehicular Technology, 2025, 74(10): 16630-16635. doi: 10.1109/TVT.2025.3572440 URL
[19]	Tong J, Guo W, Shao J, et al. Wirelessagent: large language model agents for intelligent wireless networks[J]. China Communications, 2026, 23(3): 265-285. doi: 10.23919/JCC.fa.2025-0163.202603
[20]	Chen J, Tang J, Yang X, et al. 6GAgentGym:tool use, data synthesis, and agentic learning for network management[PP/OL]. arXiv(2026-03-31)[2026-09-01]. https://arxiv.org/abs/2603.29656.
[21]	Wu X, Farooq J, Wang Y, et al. LLM-xApp: a large language model empowered radio resource management xApp for 5G O-RAN[C]// Workshop on Security and Privacy of Next-Generation Networks (FutureG) 2025. San Diego: NDSS, 2025.
[22]	Maatouk A, Ayed F, Piovesan N, et al. Teleqna: a benchmark dataset to assess large language models telecommunications knowledge[J]. IEEE Network, 2025, 40(2): 253-260. doi: 10.1109/MNET.2025.3576035 URL
[23]	Feng A, Varvarigos A, Panitsas I, et al. Telecomts: a multi-modal observability dataset for time series and language analysis[PP/OL]. v3. arXiv(2026-05-27)[2026-09-01]. https://arxiv.org/abs/2510.06063.
[24]	Ezzakri A, Piovesan N, Sana M, et al. Teletables: a benchmark for large language models in telecom table interpretation[PP/OL]. v2. arXiv(2026-09-04)[2026-09-01]. https://arxiv.org/abs/2601.04202.
[25]	Colle V, Sana M, Piovesan N, et al. Telemath: a benchmark for large language models in telecom mathematical problem solving[J]. IEEE Network, 2026:1-7.
[26]	Tong J, Liu F, Xv L, et al. WirelessBench: a tolerance-aware LLM agent benchmark for wireless network intelligence[PP/OL]. arXiv(2026-03-22)[2026-09-01]. https://arxiv.org/abs/2603.21251.
[27]	Xiao J, Lin Y, Qiu H, et al. TeleCom-Bench: how far are large language models from industrial telecommunication applications?[C]// Proceedings of the 32nd ACM SIGKDD Conference on Knowledge Discovery and Data Mining V.2. New York:Association for Computing Machinery, 2026:10056-10067.
[28]	Liu X, Yu H, Zhang H, et al. Agentbench: evaluating LLMs as agents[PP/OL]. v3. arXiv(2025-10-04)[2026-09-01]. https://arxiv.org/abs/2308.03688.
[29]	Qin Y, Liang S, Ye Y, et al. Toolllm: facilitating large language models to master 16000+ real-world APIs[PP/OL]. v2. arXiv(2023-10-03)[2026-09-01]. https://arxiv.org/abs/2307.16789.
[30]	Li Q, Chen H, Li J, et al. Dk-root: a joint data-and-knowledge-driven framework for root cause analysis of QoE degradations in mobile networks[J]. IEEE Transactions on Networking, 2026, 34:5879-5894. doi: 10.1109/TON.2026.3705395 URL
[31]	承楠, 陈芳炯, 陈文, 等. 6G全场景按需服务:愿景、技术与展望[J]. 中国科学:信息科学, 2024, 54(5):1025-1054.
[32]	Polese M, Bonati L, D’oro S, et al. Understanding O-RAN: architecture, interfaces, algorithms, security, and research challenges[J]. IEEE Communications Surveys & Tutorials, 2023, 25(2): 1376-1411.