Google Gemini 3.1 Pro Preview leaks online
Google is developing Gemini 3.1 Pro Preview as its next major AI model update, which is now leaked online. A screenshot from Artificial Analysis AI’s model selection page revealed this upcoming version. The image displays a dropdown list of available and preview models, with “Gemini 3.1 Pro Preview”. It appears alongside other unreleased Google entries, including:
- Gemini 2.0 Flash Thinking experimental (December and January variants)
- Gemini 2.0 Flash Native Audio
- Gemini 2.5 Flash Native Audio Dialog
- Gemini 2.5 Flash Native Audio
- Gemini 3 Pro Preview (high and low variants)
🚨 Gemini 3.1 Pro Preview
We are getting an update , this launch will be Google’s way to come back
Get ready for the to create new workflow for coding with ge33ini 👀 pic.twitter.com/ZfCXJYITd7
— Chetaslua (@chetaslua) February 11, 2026
The list also features new Gemma models such as Gemma 3 4B, Gemma 3n E2B, Gemma 3n F4B (including a May variant), and Gemma 7B. Competitor models like Claude 3.5 Haiku, Claude 3.5 Sonnet (June and October), and Claude 3.7 Sonnet appear for comparison.
Google launched Gemini 3 in November 2025 as its most intelligent model, emphasizing advanced reasoning for learning, building, and planning. The company followed with Gemini 3 Flash in December 2025, optimizing speed while maintaining frontier capabilities.
Gemini 3 Pro excels in instruction following, tool use, and agentic coding. Recent integrations expanded Gemini 3 into Chrome for multitasking and introduced features like Agentic Vision in Gemini 3 Flash.
Gemini 3.1 Pro Preview signals Google’s shift to faster iteration. Developers anticipate improvements in coding workflows, reduced overly compliant responses, enhanced long-context consistency, and stronger agentic performance. However, Google hasn’t announced such a release for now.
