Last week, Google announced its next-generation AI model – Gemini 2.0 Flash. It was initially available on the web version at gemini.google.com, but now is being rolled out to the Gemini app for Android. If you use Gemini on Android, you might have already seen the new model selector dropdown menu. It lets you choose the new 2.0 Flash Experimental mode.
Your Google app must be updated to version 15.50 or higher to access Gemini 2.0 Flash on Android. This model is available across all tiers, so no need for the Gemini Advanced subscription.
Tap the model switcher in the Gemini app at the top and choose between three available models:
- Gemini 1.5 Pro: Tackle complex tasks
- Gemini 1.5 Flash: Get everyday help
- Gemini 2.0 Flash Experimental: Preview our latest experimental model
Gemini 2.0 Flash is not only better than 1.5 Flash but also outperforms Gemini 1.5 Pro in benchmark, except for tasks like a novel, diagnostic long-context understanding evaluation, and automatic speech translation. This means even if you don’t have a subscription to access Gemini 1.5 Pro, you’ll still be better off with the free Gemini 2.0 Flash for most tasks.
Gemini 2.0 Flash Key Highlights
Here are some features of Gemini’s latest 2.0 Flash model:
- Gemini 2.0 Flash is a multi-modal model and can handle text, images, audio, and videos natively rather than converting them to text first.
- It can generate and modify text and images.
- It can input photos, videos, and audio recordings to answer your questions, like who said what.
- It offers quicker response times to queries and faster content creation while consuming fewer resources.
- It also demonstrates superior math, reasoning, and factuality.
- All images and audio generated by this model are watermarked with Google SynthID technology, so anyone can know these are synthetic and AI-generated. This helps avoid deepfakes and other abuse.
Gemini 2.0 Flash vs. Other Gemini Models
Here’s a quick comparison of Gemini 2.0 Flash with other Gemini models:
| Feature | Gemini 1.5 Flash | Gemini 1.5 Pro | Gemini 2.0 Flash |
|---|---|---|---|
| Speed | Fast, designed for quick responses | Slower than Flash, prioritizes accuracy | Significantly improved speed over 1.5 Flash, even faster than 1.5 Pro in some cases |
| Performance | Good for general tasks | Higher accuracy and understanding, better for complex tasks | Outperforms 1.5 Pro on some benchmarks while being faster |
| Multimodal Input | Text, Images, Video, Audio | Text, Images, Video, Audio | Text, Images, Video, Audio |
| Multimodal Output | Text | Text | Text, Images (generated), Steerable Text-to-Speech Audio |
| Tool Use | Limited | More advanced | Native tool use (Google Search, code execution, user-defined functions) |
| API | Available | Available | Available with Multimodal Live API for real-time interactions |
| Availability | Widely available | Widely available | Experimental release, wider availability in January 2025 |
The Flash 2.0 model is currently being rolled out in an experimental phase to the Gemini app on Android devices. Google says that considering its early phase, it may not always work perfectly and may make mistakes. Also, some Gemini features won’t be compatible with this model in its experimental state.
Update: Google app | Gemini app
