User Reports High Speed and Low Cost Using Gemini Flash on OpenAI Codex

A user shared their experience using a Gemini Flash model (referred to as Gemini 3.6 Flash) within the Codex AI agent, highlighting its ability to complete research and generate a web page in just 1 to 2 seconds. The user praised the model's cost-effectiveness and performance, comparing it to the "Opus 5" level. This highlights the growing integration of Google's lightweight Gemini models into developer tools like OpenAI's Codex, demonstrating the practical viability of agentic workflows for rapid content generation. It shows how developers are mixing and matching models from different providers to optimize speed and cost. The user noted that the transition from completing research to generating a promotional web page took only 1-2 seconds. However, the tweet contains potential typos regarding model names, such as "Gemini 3.6 Flash" and "Opus 5," which do not align with officially announced model versions.