Sarvam AI has made a significant leap in the field of Artificial Intelligence with the launch of its open-source foundational model supporting 10 Indian languages, including Hindi, Tamil, Telugu, Malayalam, Punjabi, Odia, Gujarati, Marathi, Kannada, and Bengali. In addition to this groundbreaking model, the company has introduced GenAI services tailored for enterprises alongside the release of Sarvam 2B, a 2-billion-parameter Indic large language model (LLM) meticulously trained on an internal dataset of 4 trillion tokens.

Revolutionizing AI Accessibility

Sarvam AI’s platform is designed to be voice-enabled and multilingual, aiming to revolutionize AI accessibility and adoption across India’s diverse linguistic and socio-economic landscape. The platform’s capabilities are poised to address the unique challenges posed by India’s vast diversity, ensuring that AI technologies are inclusive and accessible to all.

The startup’s collaboration with leading technology and industry partners has been instrumental in developing and deploying this full-stack GenAI platform. This collaboration spans multiple sectors, including Financial Services, Legal Services, Consumer Goods, Technology, Media, and Telecom, underscoring the platform’s versatility and broad applicability.

Product Suite and Offerings

Sarvam AI’s product suite includes the following:

• Sarvam Agents: These are voice-enabled, multilingual business agents that can be deployed via telephone, WhatsApp, or in-app interfaces. Available in 10 Indian languages, Sarvam Agents offers an affordable solution, starting at just ₹1 per minute.
• Sarvam 2B: As per Sarvam’s official linkedin post, It is India’s first open-source, 2-billion-parameter Indic LLM. Sarvam 2B has been trained on a massive dataset of 4 trillion tokens. It offers efficient representation for 10 Indian languages, making it a significant milestone in AI development in India.
• Shuka 1.0: As per Sarvam’s official linkedin post, This is India’s first open-source AudioLM, supporting Indian language voice input and providing text output with high accuracy. Shuka 1.0 addresses the need for precise and reliable voice-to-text conversion in multiple Indian languages.
• Sarvam Models: These best-in-class Indic models are available as APIs and include services such as translation, speech recognition, speech synthesis, and document parsing, offering robust solutions for diverse AI applications.
• A1 Workbench: A generative AI workbench designed for the legal sector, A1 features tools for regulatory chat, document drafting, redaction, and data extraction, simplifying complex legal processes and enhancing productivity.

A Vision for India’s AI Future

As per their post, Sarvam AI’s thesis for AI in India centres around four key principles:

• Voice-Enabled AI: AI should be voice-enabled and cater to the country’s diverse languages, ensuring that technology is accessible to all, regardless of linguistic or socio-economic background.
• Efficient, Small AI Models: India requires AI models that are small, efficient, and capable of being customized for specific applications. This approach ensures that AI solutions are powerful, practical, and adaptable to various needs.
• Agentic AI: AI agents that can perform tasks rather than provide information are crucial for simplifying daily life. Sarvam AI is focused on developing agents that go beyond mere interaction to deliver real value through actionable tasks.
• Sovereign AI Development: AI development in India must respect the country’s sovereignty, ensuring that data remains within national borders and that all developments adhere to local laws and regulations.

Conclusion

Sarvam AI’s launch marks a pivotal moment in India’s AI landscape. Experts believe Sarvam AI is setting a new standard for AI development in the country by focusing on inclusivity, efficiency, and sovereignty. As the company continues to innovate, it is well-positioned to play a crucial role in driving the next phase of AI adoption and integration across India, ensuring that technology serves all citizens, regardless of language or socio-economic status.

Additional sources: Economic times, the hindu business line.

Image source: Sarvam AI