Malaysia Launches ILMU: A Homegrown AI Model Trained On Local Language & Culture

The name means "Intelek Luhur Malaysia Untukmu."

Malaysia just took a big step in the AI race with the launch of its own multimodal AI model, ‘ILMU’. Prime Minister Datuk Seri Anwar Ibrahim unveiled the home-grown system at the Asean AI Malaysia Summit 2025, held at Malaysia International Trade and Exhibition Centre (MITEC).

Developed by YTL AI Labs in collaboration with Universiti Malaya, ILMU is a locally trained large language model designed to understand Malaysia’s languages, culture, and everyday context, according to Bernama. What makes it special? It’s fully developed and operated within the country, keeping data local and giving Malaysia full control over its AI future.

The name ILMU—short for Intelek Luhur Malaysia Untukmu—encapsulates its mission. ‘Intelek’ represents smart, context-aware intelligence; ‘Luhur’ stands for ethics grounded in Malaysian values; ‘Malaysia’ ensures cultural and linguistic fluency; and ‘Untukmu’ highlights that it’s built to serve all Malaysians.

According to YTL AI Labs, ILMU performs as well as top-tier models like GPT-4o and Llama 3.1. It even aced Bahasa Melayu language understanding tests, scoring higher than other leading models in MalayMMLU, which evaluates real-world knowledge in Malay.

For those eager to try it out, ILMUchat—the consumer-facing AI chatbot—will open for early access on Malaysia Day, 16 September 2025.