Tag
LemonSliceAI 推出 CWM-1 Lite，支持实时视频生成，且价格与语音模型持平，主打低成本实时视频能力。
Tavus 推出 Griffin——号称全球首个「人类交互模型」，可实时看、听、说并响应停顿、表情和手势，一次一分钟视频通话中 48% 的参与者误以为在和真人对话。Griffin-Lite 作为研究预览向早期测试者开放，支持全双工视频到视频交互及 720p 生成。
Vivix.AI launched A1, a real-time video model for interactive characters and worlds that generates audio and video together in small consecutive segments, enabling seamless interruption and live, long-form conversation with persistent state.
Benjamin DEKR shares a quote from Runway Labs about their real-time video interfaces enabling rich interactions directly in scenes.
AMAP CV Lab has introduced ABot-Recon, a real-time AI system that processes long video streams to generate 3D point clouds and camera paths, featuring stable performance and open-source code released under Apache 2.0.
Starlink-equipped buoys provided real-time video streaming in the Indian Ocean during Starship's thirteenth test flight, demonstrating Starlink's connectivity even in remote areas.
MaineCoon is a 22B real-time text-to-audio-video model that achieves up to 47.5 FPS on a single H100 GPU, enabling low-cost, long-duration streaming with synchronized speech and visuals for live AI characters.
A discussion about the evolution of AI social apps from text chat to real-time video interfaces, highlighting Mel's multimodal interaction stack and the technical challenges of latency, lip sync, and orchestration.
Elon Musk shares real-time video of Starship captured from Starlink, along with a striking photo from Flight 12.
LingBot-Map is an AI model capable of converting real-time video streams into real-time 3D reconstruction, running at 20 FPS with complete code and model provided.