Event ·Multi-source

Tavus发布Griffin实时视频对话模型

Tavus推出Griffin视频到视频模型研究预览版,可学习人类对话行为。该模型能观看并倾听对话,可在句子中打断或回应新进入画面的人。Griffin是首个通过图灵测试的模型,在120名参与者测试中45%认为对话真实。在NVIDIA视频全双工基准测试中,Griffin Lite得分为3.83,排名所有评估模型首位,接近人类3.92分。

Current read

Tavus的Griffin能实时视频对话,首个通过图灵测试,分数接近真人。

9 sources68° AI heatLast updated 2026/10/1 18:28:56

Evidence

9 sources
Primary sourceTestingCatalog
View source
Related source 2rohanpaul_ai
View source
Related source 4shao__meng
View source
Related source 6@koltregaskes
View source

Conflict check

Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.