Gemini was designed from the ground-up to be multimodal. It is highly efficient in tool and API integrations, and it is built to support future innovations like memory and planning. We're seeing multimodal capabilities that were not present in previous models.
Gemini is our most flexible model to date -- it can run on anything from data centers to smartphones. Its cutting-edge capabilities will improve the way developers and enterprises build and scale AI.
Gemini Ultra - Our largest and most capable model, designed for highly complex tasks.
Gemini Pro is our best model to scale across a variety of tasks.
Gemini Nano -- our most efficient model for on-device tasks.
Gemini Flash - our experimental model is our workhorse with low latency, enhanced performance and built to power agentic experience.