Gnani.ai Unveils 5-Billion-Parameter Voice AI Model Inya VoiceOS at AI Summit

Gnani.ai on unveiled Inya VoiceOS, a 5-billion-parameter voice-to-voice foundational AI model, at the India AI Impact Summit 2026, where the technology was inaugurated by Prime Minister Narendra Modi on Tuesday.

The Bengaluru-based company said the model has been built under the IndiaAI Mission and is being released as a research preview ahead of a planned 14-billion-parameter flagship version. Co-Founder and CEO Ganesh Gopalan demonstrated the system at the summit.

According to the company, Inya VoiceOS operates directly in acoustic and semantic space, allowing speech-to-speech interaction without relying on intermediate speech-to-text and text-to-speech systems. The architecture is designed to preserve tone, emotion, pacing and other conversational cues while enabling real-time responses.

Gnani.ai said the model has been trained on more than 14 million hours of multilingual speech pretraining data and over 1.2 million hours of task-specific speech fine-tuning data, along with more than eight trillion text tokens for linguistic grounding. It supports more than 15 Indian languages and is built to handle code-mixed speech, a common feature in Indian conversational settings.

The company stated that the system delivers sub-second latency and 24 kHz audio output, and supports streaming interactions, interruption-aware inference and overlapping speech handling.

Ganesh Gopalan said the model represents a shift from conventional cascaded voice systems. “Voice-to-voice AI is not just about faster pipelines. It is about a fundamentally different architecture that preserves what makes human conversation effective,” he said at the event.

The company said potential use cases include government helplines, grievance redressal systems and emergency response platforms, as well as enterprise deployments across banking, insurance, healthcare and logistics.

Gnani.ai added that the model has been built, trained and deployed within India to ensure data sovereignty. The company is among the firms selected under the IndiaAI Mission to develop foundational AI models for domestic use.

The announcement comes amid growing efforts by Indian technology firms and policymakers to build sovereign AI capabilities, particularly in multilingual and voice-based systems suited to India’s diverse linguistic landscape.