Google Launches 'Gemini 3.5 Flash,' Promising High-Performance AI at Lower Cost

Google is making a significant push to dominate the artificial intelligence landscape with a suite of new releases, headlined by the unveiling of its next-generation lightweight AI model, 'Gemini 3.5 Flash,' at its annual 'Google I/O' developer conference. This move signals Google's strategic intent to democratize advanced AI capabilities, making them more accessible and affordable.

The Gemini 3.5 Flash model boasts performance on par with leading AI models from Google and its competitors, but with a crucial edge: significantly faster response times and lower operational costs. Google highlighted that 3.5 Flash demonstrates improved performance in areas like agent capabilities and coding compared to its previous top model, Gemini 3.1 Pro. The company estimates that businesses could save over $1 billion annually if 80% of their daily token processing—estimated at one trillion tokens—were shifted to 3.5 Flash, a compelling proposition in an era of rising AI subscription fees.

“

Companies can significantly reduce costs by appropriately mixing Flash and other frontier models.

Beyond the core model, Google is expanding its AI ecosystem with 'Gemini Spark,' a personal AI assistant designed to work seamlessly with Google services like Gmail and Google Docs, operating 24/7 even when users are offline. Furthermore, the introduction of 'Gemini Omni' marks Google's entry into advanced video generation. This model allows users, even those without professional video editing experience, to create high-quality videos from text, voice, or image inputs, with the ability to generate realistic scenes based on physical laws. This directly challenges competitors like OpenAI, which recently scaled back its video generation services.

These releases underscore Google's commitment to innovation and its aggressive strategy to capture market share. By offering a tiered approach with models like Flash for cost-efficiency and potentially more advanced versions for specialized tasks, Google is catering to a wide range of user needs. The Hankyoreh views this as a pivotal moment in the AI race, where accessibility and cost-effectiveness, alongside performance, will be key determinants of market leadership. Google's comprehensive strategy aims to embed its AI technologies deeply into everyday digital life.

“

If 80% of the approximately 1 trillion tokens processed daily by major companies are converted to 3.5 Flash, it can save over $1 billion annually.