YTL AI Labs has launched ILMUcode, a new AI coding platform designed to give Malaysian developers access to frontier-level coding capabilities through sovereign AI infrastructure.
It was officially announced at Universiti Malaya, and ILMUcode combines an agentic AI coding assistant with ILMU-GLM-5.3, the latest addition to YTL AI Labs’ ILMU family of models, developed through its partnership with Z.ai.
The platform is designed to cover much of the software development workflow, from understanding complex codebases and writing or modifying code to troubleshooting issues and turning ideas into working software. YTL AI Labs says ILMUcode ranks among leading coding models on benchmarks including Terminal-Bench 2.1 and DeepSWE.
The launch also puts ILMUcode directly into the hands of students. Universiti Malaya will be the first university in Malaysia to introduce the platform, with 800 first-year students from its Faculty of Computer Science and Information Technology receiving RM100 in ILMUcode credits each month for three months.
That gives each participating student up to RM300 in credits to experiment with AI-assisted development, build software, and explore the capabilities of the platform.
For YTL AI Labs, ILMUcode is another step in expanding the ILMU ecosystem beyond conversational AI and into more specialized applications. The company’s flagship multimodal model currently leads the MalayMMLU benchmark for Bahasa Melayu, while the wider ILMU family also includes speech recognition, text-to-speech, and embedding models.
The company has also partnered with NVIDIA to develop its agentic ILMU-Nemo models using NVIDIA Nemotron open models, which power ILMUclaw. With Z.ai, YTL AI Labs is extending the ecosystem into software development through ILMUcode and GLM-5.3.
Developers, students, and aspiring builders can get started with ILMUcode through the platform’s website.
