Gnani.ai unveiled Inya VoiceOS at India AI Impact Summit 2026, India's first 5B parameter voice-to-v...
The AMW Read
Updates the Foundation Model player map with a non-cascaded voice-to-voice architecture specifically targeting the sovereign AI push for linguistic diversity in India.
NoveltySignificance
Foundation Models · Player MapGeopolitics
Gnani.ai unveiled Inya VoiceOS at India AI Impact Summit 2026, India's first 5B parameter voice-to-voice model that processes speech directly without ASR-TTS pipelines. Trained on 14+ million hours across 15+ Indian languages, it preserves emotional context and code-mixed conversations with sub-second latency. This end-to-end architecture marks a shift from cascaded systems and signals India's sovereign AI push for its 1.4B population's linguistic diversity. Applications span government helplines to enterprise voice workflows, with a 14B version in development.
