Indian startup Sarvam announced plans for a 1 trillion-parameter AI model, aiming to rival industry leaders OpenAI and Anthropic with cost-efficient performance breakthroughs.
For speech recognition, Sarvam has launched Saras V4, which is even better at recognizing Indic languages.
It can transcribe languages like Hindi and Bengali as well as those that other AI models can't, such as Odia, Sanskrit, and Manipuri.
The company has also improved its text-to-speech model Bulbul with Bulbul V4.
This updated version lets you add emotions like laughter or emphasize certain words for a more natural audio output.
Everything Sarvam AI announced at its Epoch Builder Edition - The Economic Times
Sarvam AI plans trillion-parameter model, sets sights on competing with ChatGPT, Gemini and Claude
Sarvam to raise $75 million from Nvidia, Glade Brook, others as part of ongoing funding round - The Economic Times
Sarvam AI launches platform to help build India-centric AI models
Sarvam AI to build trillion-parameter foundation model
Sarvam AI plans to build trillion-plus parameter model, establish San Francisco office
IBM, Sarvam partner to advance Sovereign AI adoption in India
Sarvam readying end-to-end AI stack to challenge global giants - The Economic Times
IBM, Sarvam partner to build sovereign AI solutions for Indian government, regulated sectors
Sarvam eyes next raise; Ather trims losses
Sarvam AI plans a trillion-parameter model to rival global AI leaders. The company's existing voice AI is cost-effective and…
Sarvam AI has secured $75 million from Nvidia and other investors. This investment is part of a larger $300 million funding…
