Revolutionary Krutrim: India's Answer to Global AI Giants!

Krutrim, the revolutionary AI model from India, is set to redefine the global AI landscape. Developed by Ola's CEO Bhavish Aggarwal, this multilingual AI marvel promises unmatched proficiency in Indian languages, challenging the dominance of established AI giants. With its advanced features and focus on the Indian context, Krutrim is a technological breakthrough and a testament to India's growing prowess in AI innovation.

Krutrim, hailed as "India's own AI," was recently unveiled by Ola CEO Bhavish Aggarwal. This ambitious project marks a significant entry into the burgeoning field of artificial intelligence, especially within the Indian context. Krutrim is designed as a multilingual AI model, showcasing capabilities in understanding and generating text in multiple Indian languages. This aspect differentiates it from prominent AI models like OpenAI's ChatGPT and Google's Bard.

The Krutrim project introduces two main versions of the model: the base version, named Krutrim, and a more advanced iteration, Krutrim Pro. The base model, Krutrim, is trained on 2 trillion tokens and unique datasets. It's designed to cater to the diverse linguistic landscape of India, with the capability to understand 22 Indian languages and generate text in 10 of these languages. This feature underscores its potential for wide-ranging applications across various sectors in India, where language diversity often poses a significant barrier to technology adoption.

Want to hear from the founder himself? Check out the video.

Krutrim Pro, set to be launched in the next quarter, promises to be a more sophisticated, multimodal model. This version is expected to excel in advanced problem-solving and task execution. Multimodal training implies that Krutrim Pro will integrate and process information across different types of inputs, enhancing its versatility and application scope.

At the launch event, Aggarwal demonstrated Krutrim's capabilities, highlighting its proficiency in Indian languages and context. The AI model showcased its ability to understand Indian cultural nuances and respond appropriately, a critical feature for effective communication and interaction in the Indian milieu.

Furthermore, Aggarwal emphasized the importance of building AI infrastructure alongside developing AI models. He pointed out that data centers, crucial for AI operations, consume considerable energy. In this regard, Krutrim has incorporated technology to reduce the energy cost of data centers significantly. This approach not only enhances efficiency but also aligns with broader sustainability goals.

The company aims to release an AI-focused system in package (SIP) by 2025, with prototypes expected by mid-2024. This initiative is part of a broader vision to build a homegrown supercomputer tailored for AI applications in the coming years.

The Krutrim team, which operates in India and the Bay Area (US), has developed this model in just three months. They claim that Krutrim consistently outperforms other popular AI models like GPT-4 in performance with Indic languages. The team also plans to make developer APIs available by February 2024, indicating a commitment to fostering a broader ecosystem around Krutrim.

Key Highlights of Krutrim AI

• AI Model Name: Krutrim
• Unveiled By: Ola CEO Bhavish Aggarwal
• Nature of AI: Multilingual AI model
• Primary Objective: To rival global AI models like OpenAI’s ChatGPT and Google’s Bard, with a focus on Indian languages and context
• Versions of Krutrim:
• Krutrim (Base Model): Trained on 2 trillion tokens, capable of understanding 22 Indian languages and generating text in 10 Indian languages
• Krutrim Pro (Advanced Model): Set to launch in the next quarter, a multimodal model designed for advanced problem-solving and task execution
• Key Features:
• Proficiency in Indian languages and understanding of cultural context
• Real-time coding capabilities
• Voice-enabled, with the ability to talk back
• AI Infrastructure Focus: Importance of building AI infrastructure with efficient energy use in data centers. Krutrim uses technology to reduce energy costs significantly.
• Future Plans:
• Developer APIs to be made available by February 2024
• Release of AI-focused system in package (SIP) by 2025, with prototypes by mid-2024
• Aims to build a homegrown supercomputer tailored for AI applications in the coming years
• Training and Development: The model was developed in just three months and claims to outperform popular AI models like GPT-4 in Indic languages
• Team Location: Operates in both India and the Bay Area (US)