At the India AI Impact Summit, SarvamAI introduced two cutting-edge large language models (LLMs) – Sarvam-30B and Sarvam-105B. These LLMs, trained locally in India and supporting 22 Indian languages, represent a significant advancement in AI technology. The Sarvam-30B model is designed for minimal latency and maximum efficiency, utilizing fewer tokens, while the Sarvam-105B model boasts an impressive context length of 1.28 Lakh tokens, enabling complex reasoning capabilities.
In addition to the LLMs, SarvamAI also unveiled Pravah, a token factory for industrial token production, showcasing the company’s diverse AI offerings.
The summit also witnessed strategic collaborations aimed at strengthening India’s AI infrastructure. L&T’s partnership with NVIDIA to establish a gigawatt-scale AI data center demonstrates a commitment to supporting large-scale AI workloads across various sectors. Qualcomm’s pledge of $150 Mn investments in Indian AI and tech startups, particularly in edge computing and wireless telecom sectors, underlines the growing interest and support for technological innovation in India.
These developments underscore the rising prominence of indigenous AI solutions and the increasing investments in AI technologies, positioning India as a key player in the global tech landscape.
Source: Inc42 Media