Major updates to Gemini have rolled out recently, including the launch of Gemini 3.1 Pro just yesterday. This release marks a significant leap in reasoning capabilities and “agentic” workflows.

Here are the most important news and updates as of February 2026:

Gemini 3.1 Pro (Released February 19-20, 2026)

Double the Reasoning Power: This model more than doubles the reasoning performance of Gemini 3 Pro. It achieved a verified score of 77.1% on the ARC-AGI-2 benchmark, a test focused on abstract reasoning and solving new logic patterns.
Complex Problem Solving: It is specifically designed for “complex, multi-step tasks” where simple text generation isn’t enough. It excels at synthesizing large datasets, building complex software applications, and creating advanced SVG animations.
Availability: Now available in the Gemini app and NotebookLM for Google AI Pro and Ultra subscribers. Developers can access it through the Gemini API and Google Antigravity (Google’s agentic development platform).

January 2026 Feature “Drops”

Personal Intelligence (Beta): This allows you to securely connect Gemini to your Gmail, Google Photos, YouTube, and Search. It uses your personal context (like travel plans in your emails or photos from a specific trip) to provide more tailored help.
Gemini in Chrome (Auto Browse): An agentic feature where Gemini can complete tasks for you directly in the browser—such as booking appointments or finding and purchasing tickets—rather than just giving instructions.
Agentic Vision (Gemini 3 Flash): Updated Flash to go from a passive observer to an “active investigator.” It now “explores” images to find small or blurry details, significantly reducing hallucinations in visual tasks.

Video and Multimedia Advancements

Veo 3.1 & 3.1 Fast: The video generation models received a major update in January. Key new features include:
- 9:16 Support: Native generation of vertical videos (perfect for TikTok/Shorts).
- Frame Control: You can provide the first and last frames to guide the video’s start and end points.
- Image Referencing: Use up to three images to guide the style or content of a generated video.

Ecosystem & Partnerships

Apple Integration: In January 2026, Apple officially partnered with Google to use Gemini as the foundation for the next generation of Siri.
Google I/O 2026: The annual developer conference has been scheduled for May 19–20, 2026, where even more “AI breakthroughs” are expected to be unveiled.