Google: gemini-2.5-flash-preview-04-17 API
The pioneering preview of the Gemini 2.5 family and the world's first lightweight model to support Adaptive Thinking. It broke the convention that small models struggle with complex logic, allowing developers to allocate a "Thinking Budget" for deep reasoning. Its cost-to-performance ratio in math, coding, and multi-step tasks significantly impacted the industry at launch.
- Context window: 1,048,576 tokens
- Max output: 65,536 tokens
- Input: text, image, audio, video, file
- Output: text
- Reasoning: Supported
- Tool calling: Supported
- File input: Supported
- Released: 2025-04-17
Frequently Asked Questions
What is the context window of gemini-2.5-flash-preview-04-17?
gemini-2.5-flash-preview-04-17 supports a context window of up to 1,048,576 tokens.
Does gemini-2.5-flash-preview-04-17 support function calling?
Yes. gemini-2.5-flash-preview-04-17 supports tool / function calling.
Does gemini-2.5-flash-preview-04-17 support reasoning?
Yes. gemini-2.5-flash-preview-04-17 is a reasoning-capable model.