Google’s 2.5 Flash Image takes AI editing to new leve

September 29, 2025

Google has launched Gemini Flash 2.5 Image, also known as Nano-Banana during testing, bringing advanced multi-step image editing capabilities while maintaining consistency of characters and styles. This leap gives creators greater control and precision in visual generation, with a natural flow of edits across multiple stages.

During testing, the model became a real hit, taking first place on the LM Arena Image Edit leaderboard, leaving competitors like Flux-Kontext far behind. A standout feature is its multi-turn editing capability, allowing users to layer edits without losing consistency in visual elements. It also supports image blending, combining styles and objects, and advanced transformations through simple text commands.

Another major advantage is multimodal reasoning – the model leverages world knowledge and context to make more logical choices during generation. For example, it can add contextually appropriate plants to a scene, making outputs smarter and more realistic compared to competitors.

The price is set at $0.039 per image via API and Google AI Studio, making it slightly more affordable than OpenAI’s gpt-image model and BFL’s Flux-Kontext. While AI is still not ready to completely replace Photoshop and other professional tools, Gemini Flash 2.5 Image brings us a step closer to that scenario. With its consistency of characters and preservation of details, the model has the potential to spark a creative wave similar to the Studio Ghibli phenomenon, opening the door to a new generation of viral AI-editing applications.

In brief: Tech World Highlights

OpenAI announced ChatGPT Go, a cheaper subscription tailored for India, priced under $5/month with local currency payments.
ElevenLabs introduced Chat Mode, enabling users to build text-only conversational agents alongside its voice-first systems.
DeepSeek launched its V3.1 model with a larger context window, while Chinese media attribute the delay of R2 to CEO Liang Wenfeng’s “perfectionism.”
Eight Sleep raised $100 million in funding to develop the world’s first “Sleep Agent” for proactive recovery and sleep optimization.
Runway rolled out multiple updates, including third-party model integration and visual improvements in Chat Mode.