# gemini-2.5-flash-image (nano banana) This document describes the input and output parameters of the `gemini-2.5-flash-image` model API, helping you understand the meaning of fields during API usage. --- This document only describes part of the fields used. For detailed fields of the Gemini API, see the [Gemini official documentation](https://ai.google.dev/gemini-api/docs/image-generation?hl=en). ## Request Parameters ### Request Body | Field Name | Type | Required | Default Value | Description | | -------------------------------------- | ------ | -------- | ------------------ | --------------------------------------- | | contents | array | Required | - | Content of the request, containing one or more parts. | | contents.role | string | Required | "user" | Role of the content, fixed as "user". | | contents.parts | array | Required | - | Specific parts of the content. | | contents.parts.text | string | Optional | - | Prompt text. | | contents.parts.inlineData | object | Optional | - | Inline data (e.g., images). | | contents.parts.inlineData.mimeType | string | Optional | - | MIME type of the data. | | contents.parts.inlineData.data | string | Optional | - | Base64 encoded data. | | contents.parts.fileData | object | Optional | - | File data, supports snake_case: file_data. | | contents.parts.fileData.mimeType | string | Optional | - | MIME type of the file. | | contents.parts.fileData.fileUri | string | Optional | - | URI of the file. | | generationConfig | object | Optional | - | Generation configuration. | | generationConfig.responseModalities | array | Optional | ["TEXT", "IMAGE"] | Expected response modalities, can be text or image. | ## Response Parameters | Field Name | Type | Description | | ------------------------------------------------ | --------- | ------------------------------------------- | | candidates | `array` | List of returned candidate contents. | | candidates.content | `object` | Candidate content. | | candidates.content.parts | `array` | Specific parts of the content, may include text and image data. | | candidates.content.parts.text | `string` | Text description returned by the model. | | candidates.content.parts.inlineData | `object` | Inline image data. | | candidates.content.parts.inlineData.data | `string` | Base64 encoded image data. | | candidates.content.parts.inlineData.mimeType | `string` | MIME type of the data, such as "image/png". | | candidates.content.role | `string` | Role of the content, here it is "model". | | candidates.finishReason | `string` | Reason for generation ending, e.g., "STOP". | | usageMetadata | `object` | Metadata of token usage. | | usageMetadata.candidatesTokenCount | `integer` | Tokens consumed by candidate content. | | usageMetadata.promptTokenCount | `integer` | Tokens consumed by prompts. | | usageMetadata.totalTokenCount | `integer` | Total tokens consumed. | | error | `Object` | Error information object. | | error.code | `string` | Error code. | | error.message | `string` | Error message. | | error.param | `string` | Request id. | ## Examples ### Gemini Compatible Interface We provide compatibility for the gemini `{xxx}/v1beat/models` interface. You can directly use the official SDK to call this, for example [python-genai](https://github.com/googleapis/python-genai). `POST https://api-us-ca.umodelverse.ai/v1beta/models/gemini-2.5-flash-image:generateContent` #### Image Generation (Text to Image) > ⚠️ Note: You must add responseModalities: ["TEXT", "IMAGE"] in the configuration. These models do not support image-only output. ##### ** curl ** ```bash curl --location 'https://api-us-ca.umodelverse.ai/v1beta/models/gemini-2.5-flash-image:generateContent' \ --header "x-goog-api-key: $MODELVERSE_API_KEY" \ --header 'Content-Type: application/json' \ --data '{ "contents": [ { "role": "user", "parts": [ { "text": "Create a picture of a nano banana dish in a fancy restaurant with a Gemini theme" } ] } ], "generationConfig": { "responseModalities": [ "TEXT", "IMAGE" ] } }' | jq -r '.candidates[0].content.parts[1].inlineData.data' \ | base64 -d > modelverse_generated_image.png ``` ##### ** python ** ```python from google import genai from google.genai import types from PIL import Image from io import BytesIO import os client = genai.Client( api_key=os.getenv("Modelverse_API_KEY", ""), # Your API_KEY http_options=types.HttpOptions( base_url="https://api-us-ca.umodelverse.ai" ), ) contents = [ { "text": "Create a picture of a nano banana dish in a fancy restaurant with a Gemini theme.", }, ] generation_config = types.GenerationConfig( response_modalities=["text", "image"], ) response = client.models.generate_content( model="gemini-2.5-flash-image", contents=contents, config={ "response_modalities": ["text", "image"], }, ) for part in response.candidates[0].content.parts: if part.text is not None: print(part.text) elif part.inline_data is not None: image = Image.open(BytesIO(part.inline_data.data)) image.save("modelverse_generated_image.png") ) ``` #### Image Editing (Text and Image to Image) ##### ** curl ** ```bash cat < modelverse_generated_image.png { "contents": [ { "role": "user", "parts": [ {"text": "Convert this photo to black and white, in a cartoonish style."}, { "inlineData": { "mimeType": "image/jpeg", "data": "$(curl -s https://umodelverse-inference.cn-wlcb.ufileos.com/ucloud-maxcot.jpg | base64 | tr -d '\n')" } } ] } ], "generationConfig": { "responseModalities": ["TEXT", "IMAGE"] } } EOF ``` ##### ** python ** ```python from google import genai from google.genai import types from PIL import Image from io import BytesIO import os client = genai.Client( api_key=os.getenv("Modelverse_API_KEY", ""), # Your API_KEY http_options=types.HttpOptions( base_url="https://api-us-ca.umodelverse.ai" ), ) response = client.models.generate_content( model="gemini-2.5-flash-image", contents=[ types.Content( role="user", parts=[ types.Part( text="Convert this photo to black and white, in a cartoonish style." ), types.Part( file_data=types.FileData( mime_type="image/jpeg", file_uri="https://umodelverse-inference.cn-wlcb.ufileos.com/ucloud-maxcot.jpg", ) ), ], ) ], config=types.GenerateContentConfig(response_modalities=["TEXT", "IMAGE"]), ) for part in response.candidates[0].content.parts: if part.text is not None: print(part.text) elif part.inline_data is not None: image = Image.open(BytesIO(part.inline_data.data)) image.save("modelverse_generated_image.png") ``` ### Response Example ```json { "candidates": [ { "content": { "parts": [ { "text": "That sounds incredibly unique! Here's a picture of a nano banana dish in a fancy restaurant with a Gemini theme:\n\n" }, { "inlineData": { "data": "xxx", "mimeType": "image/png" } } ], "role": "model" }, "finishReason": "STOP" } ], "usageMetadata": { "candidatesTokenCount": 1315, "candidatesTokensDetails": [ { "modality": "IMAGE", "tokenCount": 1290 }, { "modality": "TEXT", "tokenCount": 25 } ], "promptTokenCount": 16, "promptTokensDetails": [ { "modality": "TEXT", "tokenCount": 16 } ], "totalTokenCount": 1331, "trafficType": "ON_DEMAND" } } ``` ```json { "error": { "message": "error_message", "type": "error_type", "param": "request_id", "code": "error_code" } } ```