Gemini-3.1-flash-live
PulseAugur coverage of Gemini-3.1-flash-live — every cluster mentioning Gemini-3.1-flash-live across labs, papers, and developer communities, ranked by signal.
2 天有情绪数据
-
Thinking Machines 发布具有 200 毫秒处理能力的实时交互模型
Thinking Machines 发布了一类新的“交互模型”，专为实时对话式 AI 设计。这些模型以快速的 200 毫秒间隔处理音频、视频和文本，无需单独的轮次检测组件。这种架构允许连续的、交错的输入和输出流，从而能够实现边听边说以及在没有明确提示的情况下对视觉线索做出反应等功能。该系统利用两个共同训练的模型：一个用于实时对话的轻量级交互模型，以及一个用于规划和工具使用等复杂任务的后台模型，确保用户的低延迟。
-
Google tests hidden Gemini Live AI models with varied capabilities
Google appears to be testing at least seven new AI models for its Gemini Live voice assistant, as revealed by code within the Google app. These models, some with codenames like "Capybara" and "Nitrogen," offer varied ca…
-
xAI launches Grok Voice Think Fast 1.0, topping voice agent benchmarks
xAI has released Grok Voice Think Fast 1.0, a new flagship voice model designed for complex, multi-step workflows. This model excels in customer support and enterprise applications, offering low latency and high accurac…
-
Google DeepMind launches Gemini 3.1 Flash TTS, Live, and Lite models
Google DeepMind has unveiled a suite of Gemini 3.1 Flash models, including Flash TTS for advanced text-to-speech, Flash Live for real-time dialogue, and Flash-Lite for cost-efficient, high-volume workloads. These models…