MaLLaM 🌙

Multi-lingual Chat Language Model, 32k context length, Function Call, Malaysian centric and OpenAI compatible.


Able to understand Standard Malay, local Malay, Jawi, Standard English, Manglish, Mandarin and Indonesian as input.
You can communicate multiple languages in a single multi-turn session.

Natively RAG

Integrate MaLLaM 🌙 with existing knowledge base for better User-Chat experiences, we use GARIS-PANDUAN-SRF-RFP-2024-v1.pdf for an example.
If the user question is not related to the context provided, MaLLaM 🌙 knows how to politely decline to answer.

Function Call and Better Accuracy

Convert Natural text to JSON format for faster system integration, MaLLaM 🌙 achieved 100% JSON-able on NousResearch/json-mode-eval, higher is better.

Better accuracy compared to ChatGPT3.5 and open-source Malaysian Chat Models. We benchmarked on Malay test set, mesolitica/malay-llm-leaderboard, higher is better.

Developer Playground

You can play around with MaLLaM 🌙 at Dashboard

  • 350 TPS MaLLaM 🌙 Small, 500 TPS MaLLaM 🌙 Tiny
  • AI Safety Integration
  • Multiple function call functions
  • Generate code for OpenAI Python, OpenAI NodeJS and CuRL

Prepaid pricing

Best either for solo or a team.

  • Natively Multilingual
  • Natively RAG
  • Natively Function Call
  • General knowledge
  • Coding
  • Multi-turn conversation

MaLLaM 🌙 Small

0.4 USD
per 1M input tokens
1.0 USD
per 1M output tokens

MaLLaM 🌙 Tiny

0.2 USD
per 1M input tokens
0.6 USD
per 1M output tokens

Frequently asked questions

What is the different this MaLLaM 🌙 and MaLLaM 🌙 open source?

This MaLLaM 🌙 continue pretraining on more dataset, finetuned with bigger instruction dataset and aligned with human policy.

What is the rate limit?

Currently we hard limit 1M Tokens per Minute.

How to topup?

Just go to billing page and topup! Minimum 3 USD and Maximum 1000 USD.

Interested for Enterprise solution?

If you are interested to self-host in your virtual private network either on-premise or private cloud with custom solution, email us at or