Google's Astra Is Its First AI-for-Everything Agent
Astra is powered by an upgraded version of Google's Gemini model.
Alphabet, the parent company of Google, has introduced its artificial intelligence (AI) agent, Astra, powered by an upgraded version of Google's Gemini model. Revealed by CEO Sundar Pichai during the annual developer conference, Astra, touted as Google's first "AI-for-everything" agent, operates in real-time, handling queries across video, audio, and text formats.[1].
Pichai outlined plans to integrate Astra's functionalities into Google's Gemini app and broader product ecosystem this year. However, emphasizing quality control, Pichai cautioned that the rollout would prioritize integration and conversational response times.
The unveiling of Astra coincides with changes to Google's core search engine, introducing an "AI Overview" feature for US users and enhancing search capabilities with multi-step reasoning and voice/video query support.
Also, Google unveiled a suite of new AI products, including Veo for video generation, Imagen 3 for image creation, and Lyria for AI-generated music. Additionally, subscribers to Gemini Advanced will gain access to personalized chatbots called "Gems" for task-specific assistance.
Further enhancing its AI capabilities, Google has upgraded its flagship Gemini 1.5 Pro model, expanding its context window to 2 million tokens that enable better comprehension of nuanced instructions and facilitate continuity in conversations.
A day earlier, OpenAI unveiled its latest innovation in artificial intelligence (AI), introducing the world to GPT-4o just one year after GPT-4's introduction. GPT-4o is available to all, including unpaid users.
