AI Generation

Gemini 2.5 Flash LiteAPI

Our most cost-efficient multimodal model, offering the fastest performance for high-frequency, lightweight tasks. Gemini 2.5 Flash-Lite is best for high-volume classification, simple data extraction, and extremely low-latency applications where budget and speed are the primary constraints.

Test in Dashboard View Docs ← All Tools

Three ways to use this tool

REST API

One POST request. Get the result back directly — as an image, video, or JSON depending on the tool.

terminal

curl -X POST https://apiai.me/api/workflow/gemini-2-5-flash-lite \
     -H "X-API-Key: YOUR_API_KEY" \
     -F "image=@input.jpg" \
     -F "max_output_tokens=VALUE" \
     -F "negative_prompt=VALUE" \
     -F "PROMPT=VALUE" \
     -F "safety_filter_level=BLOCK_ONLY_HIGH" \
     -F "temperature=1" \
     -F "top_k=64" \
     -F "top_p=0.95"
     --output result.png

Dashboard Playground

Test this tool visually before writing any code. Upload an image, set parameters, and see the result live. When it looks right, copy the auto-generated curl command and paste it into your app.

Open the dashboard → API Toolbox
Find Gemini 2.5 Flash Lite and click it
Upload your input and adjust parameters
Copy the curl command and ship

Open Dashboard →

Batch Processing

Process hundreds of images at once without writing a loop. Upload a CSV with one row per item, set your parameters, and download the results as a ZIP when they're done.

Open the dashboard → Batches
Select Gemini 2.5 Flash Lite
Add your content and start the batch
Download results ZIP when complete

Go to Batches →

Parameters

Name	Required	Description	Default / Options
`max_output_tokens`	optional	Maximum output tokens (up to 65,536)	—
`negative_prompt`	optional	Elements or styles to exclude from the generated image	—
`PROMPT`	required	Text prompt for the Gemini model	—
`safety_filter_level`	optional	Content safety filter level	`BLOCK_LOW_AND_ABOVEBLOCK_MEDIUM_AND_ABOVEBLOCK_ONLY_HIGHBLOCK_NONE`
`temperature`	optional	Controls randomness (0 = deterministic, max 2)	`1`
`top_k`	optional	Top-K sampling: number of highest probability tokens to consider (max 64)	`64`
`top_p`	optional	Nucleus sampling: cumulative probability cutoff (0-1)	`0.95`