Gemini 3 Flash Preview (Reasoning) is a high-speed, efficiency-optimized model from Google designed to provide advanced reasoning capabilities at significantly lower latency and cost than larger frontier models. Part of the Gemini 3 family, the model is engineered for agentic workflows, complex coding assistance, and real-time interactive applications that require deep analytical processing without the overhead of massive parameter sets.
A primary feature of the model is its configurable reasoning system, which allows users to modulate the depth of internal processing via "thinking levels" ranging from minimal to high. This capability enables the model to balance between immediate responses for simple queries and prolonged internal chain-of-thought processing for PhD-level reasoning, mathematical problem-solving, and intricate technical analysis.
The model supports a 1-million-token context window and native multimodality, processing inputs across text, images, audio, video, and PDF documents. According to benchmark data, it performs competitively with previous-generation "Pro" models in specialized tasks such as coding and logical deduction while maintaining the rapid throughput characteristic of the Flash series.