Generative AI Meet ‘Kani-TTS-2’: a 400M ultimate open source text-to-speech model that runs in 3GB of VRAM with voice cloning support by February 15, 2026 February 15, 2026 The landscape of generative audio is moving toward efficiency. A new open-source contender, kani-tts-2Released by the team nineninesix.ai. This model marks a departure from bulky, … 0 FacebookTwitterPinterestEmail
