Google: Gemini 3.1 Flash Lite

Serving thousands of quick classification calls a minute, cheaply, is the workload this GA model is tuned for. Google describes it as a high-efficiency multimodal model optimised for low-latency, high-volume runs, accepting text, image, video, audio, and PDF inputs, so a concrete fit is a moderation pipeline that scans uploaded PDFs and images at scale and returns structured JSON verdicts. The 1,048,576-token context window, larger than most models listed here, still lets individual requests carry long documents, and chain-of-thought reasoning plus function calling remain available for the lightweight agentic tasks it targets. Positioning leans on price and speed: at $0.25 per million tokens it is the cheapest option in this batch, undercutting the Gemini Pro tiers, and against fuller siblings like Gemini 3.5 Flash it trades some depth for latency and cost. There is no free tier, so usage is metered from the first call. It suits engineering teams doing content-creation, research, and writing at high volume who need broad input support without premium pricing. Choose it when throughput and cost per call dominate; if a task needs the strongest reasoning on hard problems, step up to a Pro-tier model instead of this efficiency build.