Deepdub Unveils New AI Voice Model Designed for Real-Time Customer Calls
Israeli AI company Deepdub unveils Phantom Z 3.4 Conversational, a multilingual text-to-speech model for real-time enterprise voice agents, boasting improved.
>>”>Jerusalem, 6 September, 2026 (TPS-IL) — Israeli voice AI company Deepdub has launched Phantom Z 3.4 Conversational, a multilingual text-to-speech model designed for real-time enterprise voice agents.
The Tel Aviv-based company said the model delivers 150-millisecond p95 time-to-first-audio at 48 kHz and improved accuracy when reading dates, account numbers, transaction amounts and other sensitive details. Deepdub also highlighted expanded Hebrew capabilities, saying the model determines pronunciation from sentence context and ranked first for Hebrew text-to-speech on the public TTS Arena leaderboard.