Krutrim, an AI startup founded by Ola CEO Bhavish Aggarwal, is making significant progress in building a multilingual large language model (LLM). This LLM aims to generate text in 10 Indian languages, including Hindi, Marathi, Telugu, Kannada, and Odiya. The model is specifically designed to comprehend macronic text, which combines Hindi and English.
During a recent event in Bengaluru, Agarwal demonstrated the model, emphasizing that it is still a work in progress. To ensure the LLM understands the nuances of Indian languages, it has been trained on over two trillion tokens using a custom tokenizer.
A Pro version of the AI model is expected to launch in the next quarter. This advanced version will possess complex problem-solving and task execution capabilities.
Agarwal further stated, With an India-first cost structure, Krutrim will have the largest representation of Indian data, enabling us to create novel models beyond LLMs across sectors, making India the most productive, efficient, and empowered economy in the world.
The company also revealed plans to develop indigenous data centers and progress toward server computing, edge computing, and supercomputers. The targeted timeline for this project is mid-next year, with the roadmap set to be unveiled by the end of 2025.
Currently, Krutrim is running an early access program until January 2024, allowing interested users to sign up. The model is slated for release in January, and developers will have access to the beta version through APIs starting in February next year.
With its ability to generate text in multiple Indian languages, Krutrim’s multilingual large language model holds great potential for various applications. As the project advances, it aims to revolutionize the Indian economy while ensuring the seamless integration of language and technology.
Note: This content has been generated using AI language model and edited by human to adhere to the given guidelines.
