Gradium AI Releases New Default TTS Model: 81.0% Hard-Case Pass Rate at 216 ms Time-to-First-Audio

AITopTools Editorial TeamSeptember 1, 2026

What changed

Gradium AI released a new default text-to-speech model, which turns written text into spoken audio. The company reports that it passed 81.0% of 500 difficult, human-rated sentences in five languages, with the first audio arriving in a median 216 milliseconds on Coval.

What this means for you

Voice-app developers and creators may be able to deliver faster, more natural-sounding speech with this release. The 500-sentence evaluation set is available on Hugging Face under a Creative Commons license, making the claims easier to review and reuse.

Related AI tools

Explore directory listings connected to the products, companies, and workflows in this story.

Passed.ai

Related AI news

New AI featuresSep 15, 2026

Google Releases Gemini 3.8 Live and 3.8 Live Extended Thinking for Production Grade Voice Agents

The models are available now in Google’s Gemini API and AI Studio, allowing developers to build voice applications at $0.005 per minute for audio input. Generated audio includes Google DeepMind’s SynthID watermark, which identifies it as AI-created.

MarkTechPostSee why it matters

New AI featuresSep 15, 2026

Google launches Gemini 3.8 Live to take on OpenAI's GPT-Live-1 at a fraction of the cost

Developers can use Google’s models to build voice-based AI tools at a substantially lower stated cost than OpenAI’s offering. The trade-off is that GPT-Live-1 may provide more natural conversation.

The DecoderSee why it matters

Creative toolsSep 15, 2026

AI ‘Actor’ Tilly Norwood Told Me That ‘All Lives Matter’

The example shows how a promotional virtual character may handle sensitive questions, but the feed provides no evidence that Tilly Norwood is available for public use or that viewers can interact with it beyond this promotion.

WIRED AISee why it matters