ILMU (ILMUchat)
Malaysia's sovereign LLM, tuned on rojak speech, launched by the PM
Overview
ILMU is YTL AI Labs' multimodal LLM, built with Universiti Malaya and trained on Malay, Manglish and regional dialects, launched publicly by Prime Minister Anwar Ibrahim in September 2025. ILMUchat is free for individuals; businesses get token-based API access on Malaysian infrastructure.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 12 Jul 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.
What you can produce with ILMU (ILMUchat)
- ILMUchat multimodal consumer assistant (text, voice, image) — free
- LLM tuned for Malay, Manglish and rojak code-switching
- ILMU API with token-based billing and Malaysia-resident data hosting
- ILMU Claw agentic AI platform (built with Nvidia NeMo)
- Enterprise deployment case: telco Yes AI-first customer experience
- ILMU AI Accelerator Programme (up to USD 1M in API tokens for eligible teams)
ASEAN Perspective
ILMU (ILMUchat) in Southeast Asia
ASEAN-region availability and pricing notes coming soon. Drop the editorial team a note via /contact/ if you can supply local context (Singapore/Malaysia/Indonesia/Thailand/Vietnam).
ILMU's core claim — that it handles rojak, the Malay/English/Chinese/dialect code-switching Malaysians actually speak — checks out in early hands-on testing, including a review that had it give a neutral, well-sourced account of the 1MDB scandal rather than dodging it. YTL is backing this with real infrastructure: its own data centre in Johor, an Nvidia partnership for the ILMU Claw agent platform, and a live deployment inside telco Yes as of mid-2026.
The honest gap, per one hands-on review, is tone: it understands casual mixed-language prompts but tends to answer in stiff, formal Malay — described as sounding like "a government textbook," with no casual-mode toggle yet. Benchmark claims of beating GPT-4o on Malay-language tasks come from YTL's own materials rather than independent third-party testing, so treat those as a starting point, not a verdict. For sovereign-AI-in-ASEAN watchers, this is one of the more substantive builds in the region right now.
What people say
ILMU launched into public early access on Malaysia Day, 16 September 2025, with Prime Minister Anwar Ibrahim personally unveiling it as the country's first homegrown multimodal LLM — a genuine national-project framing, not just a marketing line, backed by Malaysia's broader sovereign-AI cloud push. YTL AI Labs built it with Universiti Malaya specifically to handle Malaysian code-switching: Malay, English, Chinese, Tamil and local dialects blended in a single sentence, which is how a large share of everyday Malaysian conversation actually works.
A hands-on review by Medium writer Peter Ho tested the model against a genuinely hard case — the 1MDB scandal — and found it gave a comprehensive, neutral account, citing the USD 4.5 billion figure and Najib Razak's sentencing without evasion, and referencing 33 separate sources to back its claims. That's a meaningfully harder test than most vendor demos attempt. A separate review from The Rakyat Post praised the model's grasp of "rojak" input but flagged that its output skews formal — the reviewer's phrase was that it "understands the Rojak, but it speaks like a Professor" — and suggested a casual/formal toggle would help for a product aimed at general consumers rather than civil servants.
On benchmarks, YTL and its press materials claim ILMU outperforms frontier models including GPT-4o and Llama 3.1 on Malay-language understanding, including acing PT3 and SPM national exam questions in Bahasa Melayu. These numbers come from the company rather than an independent lab, so they're worth treating as a claim pending outside verification rather than settled fact.
The product has moved from demo to real deployment quickly: YTL AI Labs announced in June 2026 that its technology now powers telco Yes as an "AI-first" customer experience layer, and in April 2026 launched ILMU Claw, an agentic AI platform built with Nvidia's NeMo framework and hosted on YTL's own Green Data Centre Park in Johor — a concrete data-sovereignty pitch for enterprise buyers wary of routing Malaysian customer data through foreign clouds. For companies wanting to build on top, YTL also runs an accelerator programme offering up to USD 1 million in ILMU API tokens to eligible participants over their first year, alongside standard token-based API pricing through the ILMU Platform.
Summary of public user & expert reviews, compiled by RECATOOLS.
About this listing
This entry was compiled from publicly available data including ILMU (ILMUchat)'s official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with ILMU (ILMUchat) unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to ILMU (ILMUchat) directly →
Spotted something out of date? Suggest an update →
ILMU (ILMUchat) in the news
Cloud & Infra
The Cloud Market Just Crossed a Half-Trillion-Dollar Run Rate — and AI Is Why It's Still A...
Privacy & Data
ASEAN's DEFA: The Region's Bet to Harmonise Cross-Border Data Rules Heads Toward a 2026 Si...
Cloud & Infra
ASEAN's Data-Centre Boom Is Hitting a Wall of Power and Water — and Johor Is Already Turni...
Alternatives to ILMU (ILMUchat)
More in LLMs & Chat
