India’s push for homegrown AI gained another boost as Bengaluru-based voice AI startup Gnani introduced 2 new platforms under its sovereign AI stack, Artha. The stack combines Evon v3.3, an open-weight foundational large language model trained from scratch, and Plexus, an agentic AI platform designed for Indian enterprises and public institutions.
The platforms were unveiled in the presence of Vice-President C P Radhakrishnan at his official residence in New Delhi on August 28. Radhakrishnan highlighted sovereign LLMs as a major advantage for India and said AI should remain open, affordable and accessible.
“Both these platforms reflect the growing strength of India’s technology ecosystem. This initiative shows that our engineers have the capability not only to use frontier technologies, but also to build them,” he said.
Evon v3.3 targets Indian languages
Evon v3.3 has 30 billion parameters, with around 3.5 billion active for each token. The model supports more than 11 Indian languages and its weights are available under the Apache 2.0 licence. Developers can request the weights for free through Hugging Face.
Gnani said Evon v3.3 uses around 20% fewer tokens per Indian-language word than the tokenizer used by the GPT-5 family and less than half the tokens required by byte-level tokenizers used by models such as DeepSeek, Llama and Qwen.
The company rebuilt its tokenizer to better handle Indian scripts and reduce token consumption. According to CEO and co-founder Ganesh Gopalan, Evon v3.3 outperformed a similarly sized unnamed model and an unnamed 105-billion-parameter Indic model on the MILU Indian-language benchmark covering 11 languages and multiple academic and professional subjects.
“Sovereign AI is not about keeping the world out. It is about India having the capability to build for itself – and then for every country that shares its problems,” Gopalan said.
Plexus brings agentic AI to enterprises
Plexus allows enterprises to create and deploy AI agents using natural language prompts. The agents can call tools and work across documents, systems and conversations. Customers can also select underlying models such as Evon v3.3.
Gnani demonstrated agents for retrieving PAN card details and deploying multi-agent workflows for welfare programmes and grievance resolution.
“Each agent is a discrete, identity-bearing unit – closer to an employee than a script – combined with other agents into workflows built around a defined outcome. These workflows run under an orchestration layer that can be human-in-the-loop or AI-led, with guardrails, observability, and audit logging built in rather than added afterwards,” Gnani said.
Gnani is among 12 entities selected under the India AI Mission to develop sovereign LLMs. The mission was approved in 2024 with an initial outlay of Rs 10,372 crore. In early 2025, the government introduced GPU subsidies and invited startups to seek compute support.
At the India AI Impact Summit in February 2026, Gnani also launched Vachana TTS, a text-to-speech model capable of cloning voices across 12 Indian languages using less than 10 seconds of reference audio. It was among 3 sovereign AI models launched on the global stage.
Also read: Viksit Workforce for a Viksit Bharat
Do Follow: The Mainstream LinkedIn | The Mainstream Facebook | The Mainstream Youtube | The Mainstream Twitter
About us:
The Mainstream is a premier platform delivering the latest updates and informed perspectives across the technology business and cyber landscape. Built on research-driven, thought leadership and original intellectual property, The Mainstream also curates summits & conferences that convene decision makers to explore how technology reshapes industries and leadership. With a growing presence in India and globally across the Middle East, Africa, ASEAN, the USA, the UK and Australia, The Mainstream carries a vision to bring the latest happenings and insights to 8.2 billion people and to place technology at the centre of conversation for leaders navigating the future.