Gemini 3.5 Flash Lite (Google Vertex AI)

Gemini 3.5 Flash-Lite is positioned for fast, economical execution where many requests must be handled quickly. Its intended workloads include document parsing, straightforward data extraction, and subagent or other high-volume agentic tasks, with multimodal inputs available when an application needs to work across text, images, video, audio, or PDF files. The source describes it as proprietary, but does not provide a parameter count, architecture details, training scale, or benchmark results.

In practice, the model suits pipelines in which response speed and operating cost matter more than a claim of frontier leadership. Its listing supports function calling, structured output, reasoning, JSON mode, streaming, fine-tuning, and batch processing, making it relevant for extraction services, document workflows, and coordinated agents. Those capabilities and the large context should be treated as catalog metadata from a third-party hub rather than a substitute for Google-official technical documentation.