Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Ola Founder Bhavish Aggarwal is investing $ 230 million into He is an AI startup established Since the country presses itself to establish itself in a case affected by the United States and Chinese companies.
Agarwal is basically funding through his family office to invest in Crutile, a source familiar with the subject has informed TechCrunch. In A post on x On Tuesday, Agarwal said Cruit was trying to attract $ 1.15 billion investment in next year. He will try to increase the rest of the capital from outside investors, the source said.
In the announcement of the fund, Unicorn Startup Crortrime made its AI models open source and unveiled the plan to create India’s largest supercomputer in partnership with the NVIDIA.
The Lab Published Cruk -2 has published a 12 -billion parameter language model that shows Indian languages ​​strongly in processing. Crutrim in the sensation analysis test Shared on TuesdayIt scored 0.95 for competitive models than 0.70, while the code generated 80% success rate in generation tasks.
The lab has unveiled a number of specialized models, including systems for images, speech translations and text search, which has been imitated for all Indian languages.
Agarwal writes, “We are still not close to the Global Benchmark but have made good progress in one year,” whose other initiatives have been supported by Softbank. To create a world -class Indian AI ecosystem. “
This initiative comes when India is trying to establish itself in an artificial intelligence landscape affected by the United States and Chinese companies. Recent DIPSC RA 1 “Argument” model publishedA clearly built on a modest budget, has sent shock waves through the technology industry.
India last week Praise the progress of the dipsek And it says that the country will host large language models of Chinese AI Lab on domestic servers. Crust Starts DEPSEC Offer to Indian Servers Last week.
Has also created the Cruit’s own evaluation structure, IndiabTo evaluate the skills of AI models in Indian language, address an interval of existing benchmarks that initially focus on English and Chinese.
The technical methods of the lab include using a 128,000-token context window, allowing its systems to handle longer text and more complex conversations. Performance Metrics, published by Startup, earned a high score on Crutrim -2 grammar correct (0.98) and multi -turn conversation (0.91).
The investment follows the introduction of a 7 -billion parameter system in January Crutrim -1, which serves as the first larger language model in India. The NVIDIA is expected to be live in March with the deployment of super computers, plans to expand throughout the year.
