Gemini 3.1 Flash-Lite: Full Overview, Pricing & Use Cases

Gemini 3.1 Flash-Lite is the most cost-efficient model in Google's Gemini 3 lineup, built for teams that need speed and scale without a heavy price tag. Where flagship models trade speed for depth, Flash-Lite flips the equation: it answers fast, costs a fraction of the price, and still handles real work — translation, transcription, data extraction, and document summarization — at a quality level that closes in on Gemini 2.5 Flash. This guide covers what Gemini 3.1 Flash-Lite actually does, how its pricing compares across the Gemini lineup, and where it fits into a Chat Smith workflow.

Gemini 3.1 Flash-Lite is Google's lightest, most cost-efficient model in the Gemini 3 series. It's built for high-volume, latency-sensitive workloads — the kind of background tasks that run thousands or millions of times a day, where every millisecond and every fraction of a cent adds up. Google rolled it out in preview through the Gemini API in Google AI Studio and for enterprise teams via Vertex AI, before making it generally available.

Compared with the older Gemini 2.0 Flash-Lite and Gemini 2.5 Flash-Lite models, this release brings a meaningful quality jump — closing much of the gap with Gemini 2.5 Flash on core capabilities like instruction following, reasoning, and audio understanding, while keeping Lite-tier pricing. It also supports four adjustable thinking levels (minimal, low, medium, high), so developers can dial reasoning depth up or down depending on how demanding a given request is.

Gemini 3.1 Flash-Lite accepts text, image, video, audio, and PDF files as input, though it only outputs text. It supports a 1,048,576-token input context window with up to 65,536 tokens of output — enough headroom to process long documents, full audio recordings, or large batches of chat history in a single call.

Not every request needs deep reasoning. Gemini 3.1 Flash-Lite lets developers set a thinking level for each call, trading a bit of latency for better accuracy on harder problems, or skipping the extra reasoning step entirely when a task is simple and speed matters more.

Beyond raw speed, the model supports batch processing, prompt caching, function calling, structured JSON outputs, code execution, and search grounding — the infrastructure agentic pipelines and automated classifiers actually need, not just a chat window.

Gemini 3.1 Flash-Lite is priced at $0.25 per million input tokens and $1.50 per million output tokens — roughly half the cost of Gemini 3 Flash. On speed, independent benchmarking from Artificial Analysis found it delivers a noticeably faster time to first token and higher output throughput than Gemini 2.5 Flash, while matching or exceeding it on quality.

Flash-Lite isn't the only model built for this tier. GPT-5 Nano and Claude Haiku 4.5 follow the same design principle — trade some reasoning depth for speed and cost — and all three are useful for the same class of high-frequency jobs. Within Google's own lineup, Gemini 3.1 Flash-Lite replaces Gemini 2.5 Flash-Lite as the default budget option, while Gemini 2.5 Flash remains the benchmark it's chasing on quality.

Chat Smith gives you direct access to Gemini 3.1 Flash-Lite alongside the rest of Google's model lineup, so you can switch models per task instead of committing to one. It's a solid default whenever you need fast, cheap output — quick replies, first-pass drafts, translations, or summarizing a long PDF — and you can always hand a task to a heavier model when the job calls for deeper reasoning.

For document-heavy workflows, pairing Flash-Lite's speed with Chat Smith's PDF summarizer makes short work of triaging long reports before you ever open the full file.

Gemini 3.1 Flash-Lite delivers most of what a flagship Gemini model offers for a fraction of the cost, which makes it a practical default for the high-volume, background work that quietly makes up most real-world AI usage. You can try it directly inside Chat Smith alongside every other model in the lineup.