Best-in-class intelligence for your high-volume workloads
Google's largest and most capable AI model. Built from the ground up to be multimodal, Gemini can generalize and seamlessly understand, operate across and combine different types of information, including text, images, audio, video and code.
The Gemini model family is documented to understand, explain, and generate high-quality code in Python, Java, C++, Go, and other languages. Flash-Lite provides this capability at a lighter compute footprint, suitable for inline code-assist features where low latency is critical.
The 'Flash-Lite' tier is designed for efficiency and scalability. Developers building consumer-scale products can use it to serve large request volumes at lower cost compared to heavier Gemini variants, while still benefiting from Google's multimodal reasoning infrastructure.
Flash-Lite is positioned as a lightweight, efficient model variant in the Gemini family. Its 'Lite' designation signals optimisation for speed and cost over peak capability, making it suitable for high-throughput API integrations such as classification, summarisation, or chat where response time and token cost are primary constraints.
Gemini models are built natively multimodal from pre-training, enabling seamless understanding of text, images, audio, and code in one request. Flash-Lite inherits this architecture, letting developers handle mixed-media inputs without stitching together separate models.