The model is divided into two versions. By default, it uses the “nothinking” version, but you can add the “-thinking” suffix to enable the model’s reasoning process.
Nothinking:gemini-2.5-flash-preview-04-17 or gemini-2.5-flash-preview-04-17-nothinking
Thinking:gemini-2.5-flash-preview-04-17-thinking
| Property | Description |
|---|---|
| Model code | models/gemini-2.5-flash-preview-04-17 |
| Supported data types | Inputs: Text, images, video, audio Output: Text |
| Token limits[*] | Input token limit: 1,048,576 Output token limit: 65,536 |
| Capabilities | Audio generation: Not supported Caching: Supported Code execution: Supported Function calling: Supported Image generation: Not supported Search: Supported Structured outputs: Supported Thinking: Supported Tuning: Not supported |
| Versions | Read the model version patterns for more details.
|
| Latest update | April 2025 |
| Knowledge cutoff | January 2025 |