Loading...
Google's family of multimodal AI models, billed per token through the Gemini API and Vertex AI.
Quick Definition
Gemini is Google's family of multimodal large language models, offered through the Gemini API and Vertex AI on Google Cloud. Like other frontier models it is priced per token, with different rates per model tier, making it a direct line item in GCP bills that FinOps teams increasingly need to track.
Gemini is Google's family of multimodal large language models, handling text, images, audio, and video. Teams access it through the Gemini API or through Vertex AI on Google Cloud, where its usage lands directly in the GCP bill alongside compute and storage.
Gemini follows token-based model pricing: separate rates for input and output tokens that vary by model tier, with long-context and multimodal requests consuming more. As with every model family, the spread between the lightweight and flagship tiers is large, so tier selection per workload is the main cost lever.
Example. A media company uses Gemini to summarize articles. Switching the bulk summarization pipeline from the flagship tier to a lighter tier, while keeping the flagship for editorial review tasks, cuts the model line item by half, and readers cannot tell the difference.
For GCP-heavy organizations, Gemini matters because AI spend now sits inside an existing cloud relationship: the same tagging, budgets, and alerts that govern infrastructure can govern model usage. The Google Gemini API pricing guide breaks down current rates and tiers, and the AI cost optimization guide shows how model spend fits a wider AI cost program.
Per token, with separate input and output rates that vary by model tier. Multimodal inputs like images and audio are converted into token equivalents.
Yes, when used through Vertex AI it is billed as part of Google Cloud, which lets existing cost allocation and budgeting tools see it.
Pick the lightest tier each task supports, trim prompts, cache repeated context where available, and set per-feature token budgets with alerts.