What’s New in Gemini Drops 2025?
Google Gemini is now a full-fledged AI ecosystem: a model family, consumer app, developer tools, coding/video/image generation, search enhancements, proactive assistant features (like calling businesses), and even robotics. With ongoing monthly feature drops and global expansion, Gemini is rapidly evolving.
- Productivity & Automation
- Agentic AI in Search
- Multimodal Features & Wearables
- API & Developer Updates
🧠 What is Gemini?
- Gemini is a family of large multimodal models developed by Google DeepMind, succeeding PaLM 2 and LaMDA.
- Originally released in Dec 2023 (Gemini 1.0 Ultra/Pro/Nano), it supports text, images, audio—even code. The current version is Gemini 2.5, launched March 25, 2025
Gemini 2.5: Two Key Variants
| Model | Best For |
|---|---|
| 2.5 Flash | Fast, everyday tasks; now the default version |
| 2.5 Pro | Complex tasks—reasoning, coding, math; supports “Deep Think” |
| 2.5 Flash‑Lite | Ultra‑lightweight, cost‑efficient variant, released June 17, 2025 |
Each natively supports text, images, audio, and offers long-context windows (up to 1 million tokens)
Major Recent Developments
- At Google I/O 2025, Google announced Gemini Drops—monthly AI feature bundles. Highlights include:
- Deep Think mode in 2.5 Pro,
- Gems integration across Gmail/Drive,
- Expansion of Gemini Live,
- Agentic Shopper,
- Enhanced veos and coding tools (Jules, Flow)
- AI Mode in Google Search now uses 2.5 Pro for deep answers and AI call-outs to businesses, plus Circle to Search for foldables and wearables.
- Gemini at IMO: Excelled at the 2025 International Mathematical Olympiad (silver→gold medal performance), solving five of six problems.
- Academic studies show Gemini 2.0 Flash Experimental is more gender-balanced than some competitors, though allows more permissive violent content—ongoing improvements are needed
Gemini Drops (July 2025)
A newly launched monthly update series, “Gemini Drops,” highlights the freshest features in the Gemini app and ecosystem
What’s New in Gemini Drops 2025?
✨ Productivity & Automation
- Gems integration expanded to Gmail, Calendar, Drive, and more—enabling smart summaries and automations (e.g., auto‑plan weekends, summarize unread emails). Available for AI Pro/Ultra users
- In the Gemini app, you can now schedule future or recurring actions, boosting planning capabilities.
📞 Agentic AI in Search
- AI-powered calling: Gemini (via Duplex) can now call local businesses—like pet groomers, pharmacies, toy stores—to check availability and prices, then send you a summary. It’s available to all U.S. users, with higher limits for Pro/Ultra subscribers.
- AI Mode / Deep Search now uses Gemini 2.5 Pro for complex reasoning, coding, math, and research-intensive queries. Pro/Ultra users get access via Search Labs.
🎥 Multimodal Features & Wearables
- Gemini Live adds real-time captions during conversations and camera-sharing support (first up on Galaxy Flip 7’s external screen)
- Veo 3 now supports 8‑second videos with audio—live in the app and API
- Extended compatibility with Wear OS 4+: interact with Gemini via voice on smartwatches (Galaxy Watch 8, Pixel, OnePlus, OPPO, Xiaomi)
- Circle to Search enhanced with AI Mode on foldables (e.g., Galaxy Z Fold7/Flip7), offering in-context follow‑up, live summaries, and gaming assistance.
🔧 API & Developer Updates
- Released in API on July 22:
•gemini-2.5-flash-lite– a super-fast, low-cost 2.5 model
•veo-3.0-generate-preview– enables video+audio generation. - Gemini Code Assist (agent-mode for multi-file tasks) now in preview for VS Code/JetBrains; core features like checkpoints, snippet context and terminal input are now GA.
Gemini as Many Products
- Gemini App & Bard Chatbot
- Free-tier: Flash plus limited access to Pro.
- Gemini Pro ($19.99/month) and Gemini Ultra ($249.99/month) unlock Deep Research, Veo video generation, Canvas, Gems, and mor
- Deep Research & Flash Thinking
- Analyze large documents and web sources into comprehensive reports; shows chain-of-thought reasoning
- Gems & Personalization
- Save custom AI “experts” (e.g., coach, code helper) with special files or instructions
- Imagen 4 & Veo 3
- From generating single images to animated videos with audio—available mainly for Pro/Ultra users
- Gemini Live & Agentic Features
- Real-time voice + visual interaction (e.g., live captions, context-aware assists).
- Circle to Search: draw & search with continuous interactions.
- AI Mode in Search: complex AI-generated answers, “Deep Search,” and even AI-powered phone calls (Duplex tech) to local shops, with summaries delivered to you
Gemini in Robotics
- Gemini Robotics (and Robotics‑ER), launched March 12, 2025, is a vision–language–action model for robots. On-device version rolled out June 24, monitored by select partners
Why It Matters
These updates mark a shift from passive assistants to proactive AI agents—Gemini now schedules tasks, automates planning, contacts businesses, generates videos, and helps code across projects. The model family is rapidly evolving:
- Gemini 2.5 family (Pro, Flash, Flash‑Lite) and Veo 3 are now fully supported across consumer apps and APIs.
- AI is becoming more integrated: from foldables and smartwatches to Search and IDEs.
✅ If You’re a User
- AI Pro/Ultra subscribers get the most enhancements: Gems, Deep Search, AI calls, Veo 3.
- Free-tier users still benefit: Circle to Search, basic AI calls, captions.
- Developers: Try the new models and Code Assist features in your IDE or Build API tools.
What’s New in Gemini Drops 2025? *The information provided above is not entirely correct. There is always the possibility of making an error when adding information. In general, we gather information from the manufacturer’s website and other