Just ahead of the holiday season, Google has introduced a set of updates to Gemini, its AI-powered assistant, highlighting the company’s efforts to remain competitive in the fast-evolving AI landscape. The updates include vibe coding tools for app creation, faster AI models, and enhanced search and translation capabilities.
One of the key features is Opal, Google’s tool for building mini AI apps called Gems, which is now integrated directly into its web browser. Opal allows users to develop multi-step apps without writing a single line of code by chaining together prompts, AI model calls, and other tools. Users can modify prompts, add features, or adjust app functionality simply by describing the changes they want to make.
A new visual editor on Gemini tracks the steps taken in app creation and allows users to preview their apps before sharing them with others. It also supports uploading images or video links for reference, helping AI generate apps from plain language prompts. Vibe coding has emerged as a growing trend in AI, with other companies, including OpenAI and Anthropic, releasing similar tools earlier this year. Swedish startup Lovable raised $200 million in early-stage funding for its own vibe coding platform.
Google has also introduced Gemini 3 Flash, a model designed for speed and intelligence. It offers two modes: a “fast” mode for quick answers and a “thinking” mode for solving complex problems. The model has improved capabilities for understanding videos and images, and can turn content into actionable plans in seconds. Gemini 3 Flash is reported to outperform Gemini 2.5 Flash and rivals larger frontier models on benchmarks such as GPQA Diamond, which tests advanced reasoning in biology, physics, and chemistry.
The company has enhanced search and translation features as well. Users in the United States can now engage in back-and-forth voice conversations through a Live With Search feature, receiving more fluid and expressive responses. Gemini is also integrated into Google Search and Translate apps, providing more natural and accurate translations. A beta version of live translation can interpret idiomatic phrases, such as “steal my thunder,” more effectively than word-by-word translation. Currently, live translation supports English to over 20 languages, including Spanish, Hindi, Japanese, and German, for users in the US and India.
These updates come shortly after Google released Gemini 3 Pro, its most powerful model for vibe coding and autonomous AI agents, and Gemini 3 Deep Think, which focuses on complex math, science, and logic problems. With the latest improvements, Google aims to enhance user experience across app development, search, and translation, positioning Gemini as a versatile AI tool for both casual users and developers.
