YTL, via its subsidiary YTL AI Labs, has introduced its locally developed large language model (LLM). Known as Ilmu 0.1, it specialises in Bahasa Malaysia (BM), the national language, and is touted to be a model that’s developed by Malaysians for Malaysians
During the National AI Office launch earlier today, YTL revealed that its LLM outperforms other models globally on the Malay MMLU (Massive Multi-task Language Understanding), a YTL AI Labs and Universiti Malaya (UM) co-developed benchmark that’s designed to evaluate AI models in BM. The company also claims that Ilmu 0.1 successfully passed both Malaysian PT3 and SPM national exams, achieving A grades in Bahasa Malaysia.
YTL envisions its Ilmu 0.1 LLM to support all industries, including finance, education, healthcare and public services, effectively in localisation. However, it is not yet available for public access, and no rollout date has been announced.
Besides introducing its in-house LLM, YTL also highlighted its RM20 billion investment in Malaysia with NVIDIA, which was announced earlier in March. With this, both companies aim to develop the country’s first AI supercomputer, powered by NVIDIA’s Grace Blackwell chips and will be housed within the former’s 500MW Green Data Centre Park in Kulai, Johor.
(Additional source: YTL AI Labs official website)
