Huan Qiu Wang Zi Xun·2025-11-19 03:54Core Insights - Google has officially launched its next-generation large model, Gemini 3, along with a new developer platform, Google Antigravity, aiming to compete with OpenAI through enhanced multimodal capabilities and reduced prompt requirements [1][2]. Group 1: Gemini 3 Features - Gemini 3 has significantly improved in understanding complex contexts, code generation, image interaction, and enterprise-level task processing, allowing users to obtain more accurate answers with fewer prompts [2]. - The model is being rolled out to subscription users and will cover various products, including Gemini applications, AI Mode search, AI Overviews, and Vertex AI, within weeks [2]. - Monthly active users for Gemini applications have reached 650 million, while AI Overviews boast 2 billion monthly active users, providing immediate traffic for the new model [2]. Group 2: Google Antigravity Platform - The Google Antigravity platform focuses on a "high-level, task-oriented" programming experience, enabling developers to quickly generate multi-step applications using natural language [2]. - It has been described as "the best Vibe coding model to date" by Google's VP of Labs, Josh Woodward [2]. Group 3: Competitive Landscape - The launch of Gemini 3 comes just 8 months after the previous version, Gemini 2.5, indicating an accelerated iteration pace [3]. - Google possesses proprietary chips, search entry points, and billions of end-users, providing a competitive edge against OpenAI's recently released GPT-5 [3]. - In the coming weeks, Gemini 3 API will be fully opened to enterprise customers through Vertex AI, with plans to expand computing power to handle peak demand during the holiday shopping season and cloud business growth [3].