Overview
Gemini is Google’s family of multimodal AI models, now deployed across enterprise and consumer products including Google Cloud, Chrome, ai-overviews, and mobile devices.
Recent Developments
-
2026-07-21: 3.6 Flash + 3.5 Flash-Lite GA; 3.5 Flash Cyber restricted dual-use via CodeMender (google-gemini-36-flash-cyber)
-
2026-07: Reported target of frozen-v2 architecture-hardwired inference silicon (google-frozen-v2-gemini-chip)
-
2026-06-24: computer-use now built-in tool in Gemini 3.5 Flash (was standalone Gemini 2.5 model) — browser, mobile, desktop agents via API (2026-06-24-google-gemini-35-flash-computer-use)
-
2026-06-12: outsider-enterprise allegedly used Gemini to generate phishing site code
-
2026-05-28: Powers ai-overviews at center of Munich court liability ruling (case 26 O 869/26)
Gemini misuse allegations originate from Google's court complaint — not independently corroborated beyond filings.
Product Variants
| Model | Use Case |
|---|---|
| Gemini Ultra | High-complexity tasks, Google Pro/Ultra subscriptions |
| Gemini Pro | Standard tasks, API access |
| Gemini Nano | On-device AI (mobile, Chrome) |
Gemini Enterprise Deployment
Merck Partnership (April 2026)
- Platform: Gemini Enterprise (Google Cloud)
- Scope: Transform Merck into “agentic AI-enabled enterprise”
- Employees: 75,000 workers alongside AI agents
- Functions: R&D, Manufacturing, Commercial, Corporate
- Value: Up to $1 billion multi-year partnership
Chrome Integration (April 2026)
- Feature: Skills for Gemini in Chrome
- Access: Side panel with slash commands
- Capabilities: Multi-tab execution, Auto Browse
- Pricing: Free basic tier, subscription for advanced features