◉GEMINI LABJP
●NANOLITE — Nano Banana 2 Lite is here: Google's fastest and most cost-efficient Gemini Image model, made for running lightweight image generation cheaply●OMNIFLASH — Gemini Omni Flash is in public preview, a natively multimodal model that lets enterprises and developers build custom, dynamic video workflows●AGENTS — Managed Agents expand with background: true for async server-side runs and polling, remote MCP server integration, and refreshing credentials across interactions●MEMORY — The Memory Bank IngestEvents API is generally available, decoupling event ingestion from memory generation so you can stream content continuously●THROUGHPUT — Provisioned Throughput now lets you submit up to seven pending orders for the same model and region●DEPRECATE — Image generation models shut down on August 17, and the Grok 4.1 family on the Gemini Enterprise Agent Platform on August 20●NANOLITE — Nano Banana 2 Lite is here: Google's fastest and most cost-efficient Gemini Image model, made for running lightweight image generation cheaply●OMNIFLASH — Gemini Omni Flash is in public preview, a natively multimodal model that lets enterprises and developers build custom, dynamic video workflows●AGENTS — Managed Agents expand with background: true for async server-side runs and polling, remote MCP server integration, and refreshing credentials across interactions●MEMORY — The Memory Bank IngestEvents API is generally available, decoupling event ingestion from memory generation so you can stream content continuously●THROUGHPUT — Provisioned Throughput now lets you submit up to seven pending orders for the same model and region●DEPRECATE — Image generation models shut down on August 17, and the Grok 4.1 family on the Gemini Enterprise Agent Platform on August 20
← Back to Blog

Gemini 3.1 Pro Is Here — What's New

gemini-3.1gemini-pronew-modelbenchmarks

The Gemini 3.1 Series

In March 2026, Google released two new models in the Gemini 3.1 series: the flagship Gemini 3.1 Pro and the cost-efficient Gemini 3.1 Flash-Lite. Coming just weeks after the initial Gemini 3 launch, this update underscores Google's accelerating pace of AI model development.

What Makes 3.1 Pro Stand Out

Major Reasoning Improvements

The headline feature of 3.1 Pro is its leap in complex problem-solving capabilities. On ARC-AGI-2, a benchmark that evaluates a model's ability to solve entirely new logic patterns, 3.1 Pro achieved a verified score of 77.1%. This represents a significant advancement in general-purpose reasoning.

Enhanced Multimodal Processing

3.1 Pro can comprehend vast datasets from massively multimodal sources, including text, audio, images, video, and entire code repositories. With a 1 million token context window, it handles long documents and complex codebases in a single pass.

How to Access

3.1 Pro is available through the Gemini API via Google AI Studio, Vertex AI for enterprise users, the Gemini consumer app, and NotebookLM.

Flash-Lite: A New Standard for Cost Efficiency

Released alongside 3.1 Pro, Gemini 3.1 Flash-Lite is optimized for high-volume developer workloads.

| Metric | 3.1 Flash-Lite | Comparison | |--------|---------------|------------| | Input cost | $0.25 / 1M tokens | — | | Output cost | $1.50 / 1M tokens | — | | Response speed | 2.5x faster (TTFAT) | vs 2.5 Flash | | Output speed | 45% faster | vs 2.5 Flash | | Context window | 1M tokens | — | | Max output | 64,000 tokens | — |

Flash-Lite maintains quality comparable to Gemini 2.5 Flash while dramatically reducing costs. It's ideal for high-throughput applications, chatbots, and data processing pipelines.

Practical Guidance for Developers

Choosing the Right Model

  • 3.1 Pro: Complex reasoning, long-form analysis, multimodal tasks, high-quality output requirements
  • 3.1 Flash-Lite: Fast responses, high-volume requests, cost optimization as a priority
  • 3 Flash: When you need better quality than Flash-Lite but don't require Pro-level reasoning

Using the API

To use 3.1 Pro through the Gemini API, specify gemini-3.1-pro as the model name. Migration from 2.5 Pro or 3 Flash is straightforward — simply update the model identifier.

import google.generativeai as genai
 
genai.configure(api_key="YOUR_API_KEY")
model = genai.GenerativeModel("gemini-3.1-pro")
response = model.generate_content("Your complex reasoning task...")

Upcoming Coverage on Gemini Lab

With the arrival of 3.1 Pro and Flash-Lite, we're planning detailed benchmark comparisons between 3.1 Pro and 2.5 Pro, cost optimization best practices with Flash-Lite, Function Calling performance analysis for 3.1 Pro, and new multimodal use cases.

The Gemini model lineup will continue to expand, and we'll keep you updated here on Gemini Lab.