Gemini 3.5 Flash Lite (Vertex AI, US)

Gemini 3.5 Flash Lite (US) extends Google's Flash-Lite line as a lightweight variant aimed at the highest-volume, lowest-margin workloads rather than frontier reasoning. Positioned as the cheapest paid tier in the Flash-Lite family, it blends multimodal input handling with tool use, structured output, and controllable temperature, so it can serve straightforward classification, extraction, and routing jobs where flagship reasoning would be wasted spend. The configuration available through Eden AI's gateway is marked GA and flagged NEW within the gemini-flash-lite series, reflecting a general-availability rollout rather than an experimental preview.

In practice, the model suits batched pipelines, simple agents, and rapid prototyping where cost per call dominates the decision more than peak intelligence. Both supplied third-party sources describe it as a value pick on a tight budget at scale, contrasting it with longer-context agentic coding models by emphasizing throughput and economy over frontier benchmarks. With a very large context window for a Lite tier and broad input modality coverage across text, image, video, audio, and PDF, it gives builders a flexible default for multimodal preprocessing, document triage, and lightweight orchestration tasks that need to stay inexpensive while remaining production-ready.