Tavus claims Griffin passes the video Turing test as 48% mistake it for a human
APPSO · wechat · 2026-10-03
Tavus unveiled Griffin, a real-time video interaction model it says is the first to pass a live "video Turing test," creating a new category: Human Interaction Model (HIM).
- In its experiment, 26 of 54 participants (48%) on a one-minute video call with Griffin-Lite believed they were talking to a human; the previous system fooled only 1 of 41 (2.4%).
- Griffin uses a full-duplex architecture: a dialogue engine reassesses the conversation sub-second, reading expressions, gaze, tone, and pauses to time responses, while an audio-video generation engine streams 720p/25fps video in 320ms chunks. It can render a full scene—face, hands, background—from one reference image.
- On Nvidia's VideoFDB benchmark, Griffin-Lite scored 3.83 in generation (human reference 3.92) vs 2.80 for the next-best Gemini 2.5 + Anam; audio-to-video latency is 0.43s on H100.
- Caveats: 54-person sample, one-minute calls, dialogue fluency scored lowest (4.9/7). Access is limited to trusted testers while the team builds AI-identity disclosure features.
More from Multimodal
- User Shares Coral Reef Creation Labeled 'Qwen 4 Max (xhigh)', an Unannounced Model Name — intellectronica · 2026-11-16
- Reddit megathread crowdsources the best open-source audio models: ASR, TTS, and text-to-music — rm-rf-rm · 2026-11-16
- Animating AI-generated 2D game art with FLUX.2 klein: masked redraws work but remain a hack — emilrueh · 2026-11-16
- Hand-editing 0.2% of weights plus 100 CPU-minutes turns an English TTS model into an 18MB Sanskrit one — kastnerkyle · 2026-11-16
- Zyphra open-sources textnorm tool, lifting TTS structured-data accuracy by up to 22 points — kastnerkyle · 2026-11-16
- AI music video tells a racing driver's comeback story — VictoryVisualsStudio · 2026-11-16