Giga AI, Inc, an India-based company that provides large language models (LLMs), has raised $3.6 million in seed funding to further its goal of helping businesses safely implement on-premises deployments of large language models.

Giga AI said in a statement on Monday that Nexus Venture Partners led the investment round.

The round also saw participation from Y Combinator, Liquid 2 Venture, and Y Combinator Chief Executive Officer and notable angel investor Garry Tan.

It is noted that strong candidates have emerged for industry adoption of open-source LLM models, and Giga ML provides a solution that enables enterprises to tailor and deploy LLMs securely on their own servers, avoiding the transmission of critical data to third-party servers. This method also offers efficiency and personalization.

The Giga ML platform allows businesses to pre-train and fine-tune a base model to meet their unique needs.

The X1 Large 32k model is the flagship product of Giga ML. It is a pre-trained and highly refined version of the llama2 70B 4K model, designed to deliver enhanced performance and capabilities.

Giga ML has already optimized the llama2 70B model to an 8k context length. This gives Giga ML a significant leg up in terms of potential and effectiveness.

The X1 Large 32k model not only succeeds in theory but also in practice, setting new standards for performance in the industry.

With a score of 8.4 on the widely used MT Bench metric for LLMs, it sets a new record for performance.

Giga ML also provides X1 code chat. These successes represent a major step forward for on-premises LLM capabilities, changing the standards by which businesses judge performance and usefulness.

Since its debut, Giga ML has garnered significant interest from a wide variety of businesses, resulting in the creation of a waiting list. Particularly interested in Giga ML’s services are businesses in the medical, legal, and financial fields.

Incorporating safety features like role-based access and code generation to boost engineering team productivity, Giga ML’s platform is now being embraced by businesses for use cases like customer service and internal knowledge search.

In the coming months, LLMs are expected to be implemented for nearly every function within enterprises, and the number of use cases is expected to grow.

Nexus Venture Partners’ Managing Director Abhishek Sharma is enthusiastic about the new cooperation with Giga ML, and he stressed the significance of safe LLM deployment in businesses. He highlighted existing difficulties, such as a lack of instrumentation and internal competence to install these models securely.

The importance of Giga ML’s contribution to the open-source LLM ecosystem was also emphasized by Garry Tan, Chief Executive Officer and President of Y Combinator.

Indian cyber security startup PingSafe launches from Stealth with $3.3M to outsmart attackers