Gemini Not Working: What’s Happening in the Last 24 Hours

Lisa Ernst · 19.02.2026 · Artificial Intelligence · 9 min

The landscape of artificial intelligence evolves at a dizzying pace, and keeping up with the latest advancements can feel like a full-time job. I have been tracking Google’s AI developments closely, and their recent strides with the Gemini models, particularly Gemini 3.1 Pro, demonstrate a clear push towards more sophisticated and integrated AI experiences.

Source: This image depicts a stylized representation of Google’s Gemini AI, with interconnected neural pathways symbolizing its advanced capabilities and broad reach across various digital platforms.

Here is a quick summary of recent developments:

• Gemini 3.1 Pro Launch: Google has introduced Gemini 3.1 Pro, a new core intelligence model designed for more complex tasks, accessible to developers and enterprise users, and to consumers with specific subscriptions.
• Gemini Enterprise Updates: Significant enhancements include the integration of Gemini 3.1 Pro (Preview), a new mobile app, and expanded data connectors, alongside critical security vulnerability fixes.
• I/O 2026 Announcement: Google’s upcoming I/O conference is set to focus heavily on AI, particularly Gemini’s role in future technologies and its potential influence on partners like Samsung.
• Interface Redesign: Google is testing a new, more organized interface for the Gemini app to manage its growing functionalities.
• AI Music Generation: Gemini now offers music generation capabilities, allowing users to create tracks across various genres with options for customization and AI watermarking.

Google’s Gemini 3.1 Pro: A New Core Intelligence

Google has significantly advanced its Gemini AI portfolio, with Gemini 3.1 Pro emerging as a central development following the Gemini 3 Deep Think update. This new model acts as the "core intelligence" behind current improvements, designed to handle more complex tasks than its predecessor, Gemini 3 Pro.

Developers can access Gemini 3.1 Pro as a preview via the Gemini API in Google AI Studio, the Gemini CLI, the Google Antigravity agent platform, and Android Studio. Enterprise users gain access through Vertex AI and Gemini Enterprise. For consumers, Gemini 3.1 Pro integrates into the Gemini app and NotebookLM, though it remains exclusive to paying subscribers with Google AI Pro or Ultra subscriptions.

Google emphasizes Gemini 3.1 Pro’s enhanced logical reasoning capabilities.

❝ Gemini 3.1 Pro achieved 77.1 percent on the ARC-AGI-2 benchmark, more than double the result of the 3 Pro version ❞

Google

Developer

This model aims to explain complex topics more structured and visually, consolidate data from diverse sources into a single overview, and foster creative projects (source). An illustrative example of its capabilities includes code-generated animations, which can create animated SVGs directly from a text prompt.

Gemini 3.1 Pro launched as a preview to facilitate real-world testing and gather feedback. Google intends to concurrently develop more extensive agent workflows before the model becomes generally available (source). Existing Gemini users will automatically receive 3.1 Pro depending on their product. In the Gemini app, Pro and Ultra subscriptions now come with increased limits. NotebookLM offers Gemini 3.1 Pro exclusively for its Pro and Ultra tiers.

Evolution of Gemini Enterprise

Gemini Enterprise, a comprehensive AI solution, has seen continuous updates, enhancing its functionality and integration across various platforms. On February 19, 2026, Gemini Enterprise expanded its capabilities with the "Gemini 3.1 Pro (Preview)" feature, replacing the earlier Gemini 3 Pro (Preview).

Key Updates for Gemini Enterprise

Administrators must activate the "Gemini 3.1 Pro (Preview)" toggle under "Model availability" to make this feature accessible to users. If "Gemini 3 Pro (Preview)" was previously enabled, "Gemini 3.1 Pro (Preview)" automatically becomes active (source).

Just a week prior, on February 12, 2026, the Gemini Enterprise Mobile App was released in Private GA (General Availability). This mobile application empowers users to interact with organizational data, agents, and execute tasks directly from their mobile devices (source). Further data connector expansions include a Notion connector in Public Preview on February 9, 2026, and a Linear connector on February 6, 2026. Gemini Enterprise also introduced welcome emails for new users on February 6, 2026, which administrators can activate or deactivate.

Security Enhancement

Notably, a security vulnerability (CVE-2026-1727) related to "Bucket Squatting" affecting Gemini Enterprise was resolved on February 6, 2026. This vulnerability could have allowed attackers to intercept, access, or manipulate data destined for a Gemini Enterprise Cloud Storage bucket. The necessary corrections were automatically deployed, requiring no user action (source).

Source: wiz.io

This image shows a vulnerability report titled "Bucket Squatting" with a highlighted alert, illustrating the security flaw affecting Gemini Enterprise that was resolved in early February 2026.

Continuous Updates

Throughout late 2025 and early 2026, Gemini Enterprise received a steady stream of updates. These included fine-tuned access control for individual Gemini Enterprise apps using App-Level IAM policies on February 2, 2026, and new actions for various data stores made available as a preview on January 30, 2026. The autocomplete function in the Gemini Enterprise chat field was deactivated on January 27, 2026, though users can still employ the "@" mention feature. Semantic similarity was added to the search functions of Gemini Enterprise and Business on January 12, 2026. Other significant additions include the Gemini 3 Flash (Preview) model on December 17, 2025, and Nano Banana Pro (Gemini 3 Pro Image) for image generation on November 21, 2025.

Google’s AI I/O 2026 and Strategic Expansion

Google has officially announced its I/O 2026 conference, with keynotes scheduled for May 19 and 20 (source). Historically, I/O events heavily feature Pixel devices, and the first Android 17 Beta, released in February, already hints at design and functional directions, including adjustments to navigation, color representation, device pairing, and potential new visual effects as foundations for future interfaces (source).

Samsung, a major player in the South Korean economy, is expected to incorporate these impulses into its next One UI generation. Gemini, Google’s AI platform, is poised to play a central role at I/O 2026. New interfaces, deeper app integrations, and more context-aware functions from Gemini could directly influence Samsung’s Galaxy AI strategy, especially concerning its next generation of devices.

Interface Overhaul and Music Generation

Upcoming Interface Redesign

The increasing functionality of Google Gemini has led to a more packed interface. Google is actively testing a new design for the attachment area to enhance clarity. This new interface aims to replace the current design with four attachment buttons. It introduces larger icons for essential tools and a second row for horizontal scrolling to access additional features (source). These adjustments will apply directly within the Gemini app and other applications that utilize the attachment page.

While the new interface is still under development and not widely distributed, a screenshot from Android Authority showcases the updated layout, placing Drive, NotebookLM, and Map in a second row beneath core functions like photos, camera, and files.

Source: androidauthority.com

A screenshot from Android Authority illustrates the new Gemini app interface, featuring larger icons and a second row for Drive, NotebookLM, and Map below core functions.

AI-Powered Music Generation

Beyond interface improvements, Google Gemini gained music generation capabilities on February 19, 2026. Users can select from 16 different genres, including K-Pop, 90s-Rap, or Latin Pop, and even humorous categories like "Bad Music" and "Birthday-Roast." Music generation begins by starting a new chat and selecting the "Create Music" button or the plus symbol.

Users can prompt Gemini with a theme without needing to provide their own lyrics. Opting for "cinema effect" produces an instrumental track. However, song lyrics may not always be generated in German if not explicitly requested in the prompt, leading to English 90s-Rap or Spanish Latin Pop, for instance.

While users can influence parameters like style, rhythm, or voice, creative control over the final song remains limited, with Gemini ultimately determining the outcome, which can vary even with identical prompts. A notable feature allows users to upload a photo or video, and Gemini will generate a musical piece that matches the mood. Lyria 3, the underlying model for this feature, is also integrated into YouTube’s Dream Track. The free version currently allows for a maximum of 30 seconds of music generation, with longer limits anticipated for paid AI subscriptions (AI Plus, Pro, or Ultra).

AI-generated music from Gemini includes an invisible watermark called SynthID for identification. To verify a song, users can upload it to the latest version of Gemini and instruct the AI to check it. This verification process also extends to AI-generated images and videos from Gemini.

Conclusion

Google’s continued investment in its Gemini AI models, particularly the advancements with Gemini 3.1 Pro, signifies a strategic push towards more sophisticated and integrated AI solutions across developer, enterprise, and consumer sectors. The constant stream of updates to Gemini Enterprise, coupled with interface redesigns and innovative features like AI-powered music generation, points to a future where AI becomes an increasingly seamless and powerful tool in daily digital interactions. The upcoming I/O 2026 conference will likely reveal further ambitious plans, reinforcing Gemini’s central role in Google’s ecosystem and its influence on future AI-driven technologies.

Frequently Asked Questions