Post-Cutoff.com
  1. Home
  2. Timeline
  3. 2026
  4. Tavus unveils Griffin, a real-time face-to-face video…

Tavus unveils Griffin, a real-time face-to-face video model it says is the first to pass a 'video Turing test' (48% thought it was human)

★★★after cutoffmedia-generationTavusconfidence: medium

Tavus announced Griffin, a full-duplex video-to-video "Human Interaction Model" that watches and listens while it talks, can interrupt and be interrupted, and reacts to gestures and objects on camera. In Tavus's own blind study, 48% of 54 participants who had a one-minute live video call with it thought they had talked to a real person (its earlier systems: at most 2%). It is a research preview, not a general release. The launch post reached about 5.4M views on X.

Key facts

What happened

Tavus, which makes conversational video avatars, introduced Griffin as a new model class for live face-to-face conversation: it perceives video and audio continuously and generates a talking video response in real time, handling overlap, interruptions and visual context (e.g. coaching a Rubik's cube solve). The headline claim comes from Tavus's own study; the VideoFDB numbers come from NVIDIA's benchmark leaderboard.

Why it matters

Real-time video avatars that half of short-call participants mistake for people move deepfake-style capability into live, interactive settings: customer service and tutoring, but also fraud and impersonation on video calls. The evidence so far is vendor-reported and based on one-minute calls.

Changelog

  • 2026-10-02: created (reported by the site owner; 5.4M-view launch post)
  • 2026-10-02: evening sweep: added CEO post (3.35M views), safety-gating post, and Roose/Mostaque reactions
  • 2026-10-03: morning sweep: added the Oct 2 backlash (Walsh, IterIntellectus, Sanders) and Raza's Oct 3 statement on safeguards and disclosure
  • 2026-10-03: added the reach of Tavus's follow-up launch-thread posts (sweep 2026-10-03)

Models

Videos (1)

48% of People Thought This AI Was a Real Human | Introducing Griffin

Tavus · 2026-10-01 · official

Description by Gemini, which watched the video:

Summary This official launch video from Tavus introduces "Griffin," an interactive human-like AI video model. Tavus co-founder Hassaan Raza engages in a video call with a photorealistic AI conversational avatar that demonstrates visual perception, spontaneous gesturing, and natural conversational banter.

What is shown

  • Conversational video call UI: Hassaan opens his laptop and begins a video call with an interactive AI persona via a web application interface [00:03 - 00:36].
  • Dynamic gesture generation: Hassaan asks the avatar for "just a thumbs up," and the avatar naturally generates the gesture on screen [00:38 - 00:43].
  • Visual multi-person awareness: A colleague walks up behind Hassaan; the avatar notices him visually and quips, "Oh, who's that with the ugly sweater?", reacting to his response [00:51 - 01:01].
  • Contextual and environmental awareness: The avatar notices the laptop battery is at 1% and quips that she will slow down her breathing to conserve CPU, visually demonstrating slower breathing motions [01:02 - 01:12].
  • Tavus end card: Concluding title screen showing the Tavus logo and the tagline "the human computing company" [01:41 - 01:48].

Claims & numbers

  • The video title claims "48% of People Thought This AI Was a Real Human" (referencing a video Turing test).
  • Hassaan states that it feels like "passing the Turing Test soon" [01:23].
  • No specific technical metrics, inference latency numbers, or pricing details are explicitly cited in the video audio.

Notable quotes

  • [00:28] "Do you think they'll believe that you're AI?" / "I think that's the whole point, isn't it? If it's good enough, they won't even have to think about it."
  • [00:51] "Oh, who's that with the ugly sweater?"
  • [01:14] "It's like the line between talking to a machine and just talking to someone is getting really, really thin."

Assessment This is an official promotional product demo showcasing Tavus's Griffin model capabilities in conversational video generation and visual input perception. While the interaction demonstrates seamless multimodal conversational capabilities, the presentation is a produced, edited video rather than a raw, continuous live-stream test with latency telemetry.

Described by gemini-3.8-flash on 2026-10-02 from the video's audio and frames.

People

Bernie Sanders Emad Mostaque Kevin Roose

Related posts (10)

Related events

  1. Alan Turing proposes the 'imitation game' (Turing test) ★★★★★

Sources (15)

id: 2026-10-01-tavus-griffin-video-turing-test · updated 2026-10-03 · open in the interactive timeline