See what 4 builders are making with Gemini 3.8 Flash

Google has unveiled Gemini 3.8 Flash, a cutting-edge AI model engineered for intricate reasoning and software engineering tasks. Builders are leveraging this powerful tool to craft a diverse range of applications, from simulating satellites to generating immersive 3D engineering models.

Gemini 3.8 Flash is engineered to excel in complex problem-solving, demonstrating remarkable prowess in areas such as software engineering, agentic tasks, and specialized multistep reasoning. Unlike its predecessor, 3.7 Flash, 3.8 Flash works diligently on challenging tasks, executing additional reasoning steps and invoking tools iteratively, thereby delivering more precise results.

Since its launch, builders have been actively utilizing Gemini 3.8 Flash to create innovative projects. Notable examples include Ashutosh Shrivastava's creation of satellite paths using Gemini 3.8 Flash and Google Antigravity, as well as Noctus's creation of a dynamic ink painting inspired by traditional Japanese Seigaiha waves.

The model's performance gains are particularly pronounced in multistep reasoning within STEM and other domains. For instance, Emily challenged Gemini 3.8 Flash with the task of constructing a T. rex skeleton, specifying stringent requirements, prohibited shortcuts, and accuracy checks. Hakm tackled an engineering challenge by prompting Gemini 3.8 Flash to develop an interactive model of an automatic transmission, complete with multiple camera views, display modes, smart labels, and a side panel providing detailed information on transmission components.

Gemini 3.8 Flash has demonstrated significant improvements over 3.7 Flash, often matching the performance of higher-cost frontier models. Builders can access and experiment with this powerful AI across Google's suite of products and tools, including Google Antigravity and Google AI Studio.