The Rundown AI homepage

Independent tool overview

Gemini 3.1 Pro at a glance

Gemini 3.1 Pro is Google's advanced reasoning model for complex coding, research, analysis, and agentic workflows, with multimodal inputs and a one-million-token context window.

Visit the official Gemini 3.1 Pro site ↗
Gemini 3.1 Pro product preview
Model ID
gemini-3.1-pro-preview
Status
Preview
Input context
1,048,576 tokens
Maximum output
65,536 tokens
Inputs
Text, image, video, audio, and PDF
Output
Text
Knowledge cutoff
January 2025, with grounding tools available
Last reviewed
August 29, 2026

Overview

What Gemini 3.1 Pro is

Gemini 3.1 Pro is the reasoning-focused model in Google's Gemini 3 family. It is intended for work where a fast answer is less important than following a long chain of requirements, combining information from several sources, using tools, or reasoning across a large codebase or document set.

The API accepts text, images, video, audio, and PDFs and returns text. It supports function calling, structured output, code execution, context caching, Google Search grounding, URL context, Google Maps grounding, and adjustable thinking levels. Google also exposes a custom-tools endpoint for agentic coding workflows that mix shell commands with developer-defined tools.

Consumers can use 3.1 Pro in the Gemini app and NotebookLM, while developers can access the preview model through Google AI Studio, the Gemini API, Vertex AI, Gemini CLI, Android Studio, and Antigravity. The API endpoint remains a preview model, so production systems should isolate it behind versioned configuration and regression tests.

Use cases

Who Gemini 3.1 Pro is best for

The strongest fit depends on the job you need the product to complete, not the size of its feature list.

Large-codebase reasoning

Analyze repositories, plan multi-file changes, debug difficult behavior, and coordinate tool-based implementation work.

Complex research synthesis

Combine long documents, URLs, search results, tables, and multimodal evidence into a structured analysis.

Agentic workflows

Use function calling, structured outputs, code execution, and custom tools for reliable multi-step tasks.

Multimodal document analysis

Reason across PDFs, images, recorded audio, and video when the answer depends on more than plain text.

Hard planning problems

Work through long requirement sets, tradeoffs, constraints, and dependencies before producing a decision or implementation plan.

Capabilities

Core Gemini 3.1 Pro features

1

One-million-token context

Process large codebases, document collections, long recordings, or extended conversations in a single request when the application supplies the relevant material.

2

Adjustable thinking

Choose low, medium, or high thinking levels to balance latency and reasoning depth; high is the default for 3.1 Pro.

3

Multimodal understanding

Accept text, images, video, audio, and PDF inputs while producing text responses for analysis and downstream automation.

4

Grounding and context tools

Use Google Search, Google Maps, URLs, file search in AI Studio, and code execution to bring external or computed evidence into a response.

5

Function calling and structured output

Connect business actions and developer tools, then constrain responses to a defined schema for more reliable automation.

6

Custom-tools endpoint

Use gemini-3.1-pro-preview-customtools when an agent must prioritize shell and developer-defined tools, with the tradeoff that quality may fluctuate outside those workflows.

Process

How the Gemini 3.1 Pro workflow works

  1. Step 1

    Choose the right access surface

    Use the Gemini app or NotebookLM for interactive work, AI Studio for prototyping, and the Gemini API or Vertex AI for a production integration.

  2. Step 2

    Define the output contract

    State the goal, constraints, audience, acceptable sources, and required structure; use a schema when another system will consume the answer.

  3. Step 3

    Supply only relevant context

    A large context window is capacity, not a requirement. Curate files and sections so cost and attention are spent on evidence that affects the result.

  4. Step 4

    Select tools and thinking level

    Enable search, URLs, code execution, or functions only where needed, and lower thinking depth for routine requests that do not justify Pro-level latency and cost.

  5. Step 5

    Validate important outputs

    Check citations, calculations, code, tool arguments, and factual claims before acting, especially when the model is allowed to change external systems.

  6. Step 6

    Control cost and model drift

    Use context caching or Batch/Flex options where appropriate, log token usage, pin the endpoint, and rerun evaluation cases before adopting a new preview revision.

Cost

Gemini 3.1 Pro pricing and free plan

Gemini 3.1 Pro has consumer subscription access and usage-based developer pricing. API rates increase when a prompt exceeds 200,000 tokens. Consumer availability and quotas are compute-based and vary by prompt complexity, features, conversation length, plan, and region.

Gemini Free

$0

Personal Gemini access with varying availability for 3.1 Pro.

  • Limited and compute-based 3.1 Pro access
  • 15GB Google storage
  • Gemini app
  • Deep Research and multimodal features with limits

Google AI Pro

$19.99/month

Higher consumer access to Google's Pro models and productivity features.

  • 4x Free usage access in the Gemini app
  • 5TB Google storage
  • Gemini in Gmail, Docs, Vids, and more
  • Expanded Notebook and Flow access

Google AI Ultra 5x

$99.99/month

Premium personal plan with five times the Pro usage limits.

  • Higher access to 3.1 Pro
  • Advanced feature access
  • 10,000 Flow credits
  • Cloud storage starting at 20TB

Google AI Ultra 20x

$199.99/month

Highest listed personal usage tier.

  • 20x Pro usage limits
  • Highest access to Pro models
  • 25,000 Flow credits
  • Priority access to advanced features

Gemini API — prompts up to 200K

$2 input / $12 output per 1M tokens

Standard API pricing for shorter contexts.

  • Text, image, and video input use the listed token rate
  • Thinking tokens count toward usage
  • Caching and alternate consumption options supported
  • AI Studio prototyping available

Gemini API — prompts over 200K

$4 input / $18 output per 1M tokens

Higher rate applied to requests with large input contexts.

  • Applies above 200,000 input tokens
  • Large-context output is billed at the higher output rate
  • Monitor repeated conversation history
  • Consider caching and retrieval before sending full corpora

Vertex AI / Enterprise

Usage-based or contracted

Google Cloud access with enterprise deployment, governance, and billing options.

  • Preview model in Model Garden
  • Enterprise security and administration
  • Regional availability and quotas apply
  • Cloud pricing and terms may differ from the Gemini Developer API

Pricing checked . Check current pricing at the source ↗

Assessment

Gemini 3.1 Pro strengths and limitations

Where it stands out

  • Strong fit for difficult reasoning, coding, research, and multi-step tool use
  • One-million-token context can accommodate large repositories and document sets
  • Native multimodal input reduces the need for separate transcription and extraction pipelines
  • Broad tool support includes search grounding, URLs, Maps, code execution, function calling, and structured output
  • Available across consumer, developer, coding, and enterprise Google products

What to consider

  • The API model is still in preview, so behavior, quotas, endpoints, and availability can change before general availability
  • It returns text rather than generating images or audio and does not support the Gemini Live API
  • Requests above 200,000 input tokens move into a more expensive pricing tier
  • High thinking can increase time to first answer and total billed output tokens
  • The large context window does not guarantee that every supplied detail will receive equal attention
  • Grounding reduces but does not eliminate factual errors; citations and tool results still require verification
  • Consumer usage limits are compute-based rather than a guaranteed fixed number of prompts

Compare

Gemini 3.1 Pro alternatives

The right alternative depends on the specific output, workflow, controls and budget your project requires.

Business Operations

ChatGPT

Choose ChatGPT for OpenAI's broad consumer ecosystem, custom tools, multimodal creation, and integrated agent experiences.

Explore ChatGPT

Project Management

Claude

Choose Claude for Anthropic's long-context writing, coding, analysis, and agent workflows, especially if you prefer its product ecosystem.

Explore Claude

Business Operations

DeepSeek

Choose DeepSeek when open-model availability and lower-cost reasoning options are more important than Google's integrated tools and multimodal stack.

Explore DeepSeek

Questions

Gemini 3.1 Pro FAQs

What is Gemini 3.1 Pro best used for?

It is designed for complex reasoning, software engineering, research synthesis, multimodal analysis, and agentic workflows that require reliable tool use across multiple steps.

Is Gemini 3.1 Pro generally available?

No. The Gemini API endpoint is still named gemini-3.1-pro-preview, and Google lists all Gemini 3 API models as preview. No shutdown date has been announced for this endpoint.

How large is the Gemini 3.1 Pro context window?

The model supports up to 1,048,576 input tokens and 65,536 output tokens. Sending more than 200,000 input tokens uses the higher API price tier.

Can Gemini 3.1 Pro analyze video and audio?

Yes. It accepts video and audio alongside text, images, and PDFs. Its output is text; audio generation, image generation, and the Live API are not supported by this model.

How much does the Gemini 3.1 Pro API cost?

For prompts up to 200,000 tokens, Google lists $2 per million input tokens and $12 per million output tokens. Above 200,000 input tokens, rates rise to $4 input and $18 output per million tokens.

Can I use Gemini 3.1 Pro for free?

The Gemini app includes varying free access to 3.1 Pro. API prototyping may also have free allowances in Google AI Studio, but availability, limits, and supported features can differ from paid API access.

What is the custom-tools endpoint?

gemini-3.1-pro-preview-customtools is a separate API endpoint optimized to prioritize shell commands and developer-defined tools in agentic workflows. Google notes that quality may fluctuate on tasks that do not benefit from those tools.

Bottom line

Our Gemini 3.1 Pro verdict

Gemini 3.1 Pro is a compelling choice when a job combines long context, multimodal evidence, difficult reasoning, and tools. It is overkill for simple chat or high-volume classification, where a current Flash model will usually be faster and cheaper. For production agents, the key caution is its preview status: pin the model, measure real task performance, cap spend, and keep human or programmatic checks around consequential actions.

Visit Gemini 3.1 Pro website ↗
The Rundown University

AI training for the future of work.

Get access to all our AI courses, hundreds of real-world AI use cases, live expert-led workshops, an exclusive network of AI early adopters, and more.

AI Courses

Get unlimited access to all of our current & upcoming industry-specific AI courses for the duration of your subscription.

Daily Guides

To keep up with the rapid pace of AI, our team publishes AI implementation guides daily. Our library contains 300+ practical use cases to automate real-world work.

Workshops

Join weekly, live, interactive sessions with industry leaders who are at the forefront of AI for hands-on implementation guidance and exclusive insights.

Community

Network with an exclusive community of AI-first professionals who are working smarter with AI. Learn how early adopters are using AI in their work and businesses.