Tavus 推出视频到视频模型 Griffin 的研究预览版，用于实时面对面对话，可在对话中打断或对入镜者作出反应。Tavus 称 120 名参与者中 45% 认为其对话对象是真人；在 NVIDIA Video Full-Duplex Benchmark 上 Griffin Lite 以 3.83/5 排名第一（真人为 3.92）。完整版 Griffin 尚未发布，Tavus 正在评估其发布政策。
Tavus has introduced Griffin, a research preview of its newest video-to-video model for real-time face-to-face conversations.
Griffin is designed to learn conversational behavior from people. It can watch and listen during conversations, allowing it to interrupt mid-sentence or respond when someone enters the frame.
> Tavus says Griffin is the first model to pass the Turing test, with 45% of 120 study participants taking its conversation as real.
> On NVIDIA's Video Full-Duplex Benchmark, Griffin Lite ranks first among every model evaluated, at 3.83 out of 5 against 3.92 for real humans.
> The full Griffin model remains unreleased while Tavus reviews its release policy.
Introducing Griffin, the first model to pass the video Turing test. 48% of people who talked to it live thought it was a real human. Previous systems have had a pass rate <3%. It is #1 on NVIDIA's benchmark for full-duplex AI video. It’s the first Human Interaction Model (HIM).在 X 查看被引用的帖子
来源：🚨 AI News | TestingCatalog · x.com