ADM+S RMIT team win first prize at international RAG challenge
Author ADM+S Centre
Date 11 December 2025
Congratulations to the team of ADM+S Researchers from RMIT, who have won first place at the Massive Multi-Modal User-Centric Retrieval-Augmented Generation (MMU-RAG) Challenge at NeurIPS Conference 2025. The team consisted of several ADM+S researchers, including:
- Kun Ran
- Oleg Zendel
- Damiano Spina Valenti
- Shuoqi Sun
- Khoi Vo Nguyen
- Nguyễn Đình Anh Khôi (RMIT)
- Chenglong Ma
- Marwah Alaofi
- Sachin Pathiyan Cherumanal (ADM+S Alumni)
- Danula Hettiachchi
- Falk Scholer
- Lida Rashidi (RMIT)
The inaugural MMU-RAG competition took place at the 39th edition of the Annual Conference on Neural Information Processing Systems (NeurIPS 2025).
“Our team, comprising RMIT students and staff, placed first in the open-source systems track under the dynamic user-based evaluation,” said Oleg Zendel.
“This evaluation used a chatbot arena format where users submitted any query they wanted and compared the responses from several systems side by side.”
Out of 81 total registered teams, just 8 managed to submit a fully working system, due to the challenging technical requirements.
The MMU-RAG challenge is a new international competition, developed by Carnegie Mellon University’s Language Technologies Institute (LTI) in partnership with Amazon.
It was launched to evaluate the next generation of Retrieval-Augmented Generation (RAG) systems by recreating the complexity of real-world information needs. RAG systems combine large-scale information retrieval with AI text generation, allowing them to produce informed and contextually relevant responses.
These systems are increasingly used in applications like advanced chatbots, digital assistants, and research tools.
This marks the teams second win this year in a RAG competition, following on from their win of the LiveRAG competition at the 2025 SIGIR conference.
