Nano Banana Pro Powers Image Generation
Summary
The rapid iteration of Gemini 3 capabilities, from image generation to security hardening, signals a shift in how developers will build consumer AI applications.
- Gemini 3 Image Model Debut Google DeepMind officially launched Nano Banana Pro, its Gemini 3 Pro Image model, available for developers [7, 11].
- AI Jailbreak Mechanism Found Researchers detailed ‘Adversarial poetry’ as a universal single-turn mechanism to jailbreak LLMs, submitted Nov 19 4.
- Open Source AI Path Set AI2 released Olmo 3, providing full model flow and traceability, aiming to lead open-source AI efforts 6.
- New Image App Launched Mixup, an iOS app using the Nano Banana model, debuted, allowing AI image creation from photos and text 2.
- AI Verification Integrated Gemini app gains AI image verification capabilities to combat misuse, detailed Nov 20 10.
- Nov 20, 2025 - Date of Nano Banana Pro model announcement by Google DeepMind 11.
- 19 Nov 2025 - Submission date for the paper detailing LLM adversarial poetry jailbreak 4.
- v1.4.2 - DuckDB version detailed alongside new Data-at-Rest Encryption feature 5.
Key Moments
-
Build with Nano Banana Pro, our Gemini 3 Pro Image model
— Article [7] -
Adversarial poetry as a universal single-turn jailbreak mechanism in LLMs
— Article [4] -
Olmo 3: Charting a path through the model flow to lead open-source AI
— Article [6] -
Gemini refused to believe it was 2025, and hilarity ensued
— Article [1] -
How we’re bringing AI image verification to the Gemini app
— Article [10]
Different Perspectives
All Articles
-
[11] Nano Banana Pro