Early this morning.Google formally presented its state-of-the-art image generation and editing model, Gemini 2.5 Flash Image (a.k.a. nano banana).
According to the official introduction, the main features of Gemini 2.5 Flash Image include the following:
Full character consistency: it's easy to place the same character in different environments or show the same product from multiple perspectives, while perfectly keeping its core subject intact.
Cue-based image editing: allows users to make precise local modifications to images through simple natural language commands .
Leverage Gemini's real-world knowledge: Models can leverage Gemini's powerful world knowledge base to make image generation "smarter".
Multi-image fusion: objects from one image can be 'put' into the scene of another image, and the whole process can be done with a single prompted command.
In terms of performance, Gemini 2.5 Flash Image was ranked first in several benchmarks, outperforming models such as OpenAI ChatGPT 4o (GPT Image 1 high) and Qwen Image Edit.
With regard to calling the API, the specific pricing is $30 per million output tokens, and the official introduction, generating a picture consumes about 1290 output tokens, that is to say, the cost of each picture is about $0.039, which is less than 3 cents in RMB.
Gemini 2.5 Flash Image is now accessible through the Gemini APP, Gemini API, Google AI Studio and Vertex AI.