Gemini 3.1 Pro Preview sits inside the Gemini 3 generation of Google's frontier lineup, framed as the paid preview tier for long-context reasoning, multimodal input, and tool-heavy developer work. A third-party setup guide built on Google's own model, pricing, rate-limits, API-key, compatibility, and developer pages positions it as the current default path in the Gemini Pro family, superseding the earlier Gemini 3 Pro Preview identifier, which is reported to now redirect to the new model. A companion endpoint, gemini-3.1-pro-preview-customtools, is documented for agent workloads that mix shell execution with registered tools, signaling a focus on programmable, multi-step pipelines rather than single-turn chat.

In practical terms, the model is aimed at developers who need a single endpoint that can ingest large prompts and route reasoning through external tools, with batch processing available at half the listed rates for cost-sensitive runs. The same guide notes staged pricing that rises once prompts cross a 200k-token threshold, which is a useful signal that very long contexts are an intended use case rather than an afterthought, and that prompt size should be factored into budget planning. For teams already invested in Google's ecosystem who want a preview-grade model with extended reasoning and tool-calling support before committing to a generally available release, this Preview tier offers a straightforward on-ramp through Google AI Studio and the standard Gemini API surface.