Gemini 2.5 Flash-Lite
2025-06-17 · Publisher : Google · Domain : Artificial Intelligence
Google launches Gemini 2.5 Flash-Lite, the cheapest and fastest model in the 2.5 family: preview on June 17, 2025, stable version on July 22 at 0.10 USD for input and 0.40 USD for output.
What changes
- Improvement over 1.5 and 2.0 Flash on most evaluations, with lower latency and higher decoding throughput.
- Reasoning model whose thinking is disabled by default (unlike the other 2.5 models), and can be enabled through a parameter.
- Support for Google Search, code execution, URL context and function calling.
- July 22, 2025: stable version, audio input price reduced by 40% since the preview.
### Specifications - Identifiers: `gemini-2.5-flash-lite-preview-06-17` then `gemini-2.5-flash-lite` (stable). - Context: 1,048,576 input tokens, 65,536 output tokens; knowledge cutoff: January 2025. - Inputs: text, image, video, audio, PDF. - Availability: Gemini API, Vertex AI. ### Pricing | Model | Input (USD / 1M tokens) | Output (USD / 1M tokens) | |---|---|---| | Gemini 2.5 Flash-Lite (stable) | 0.10 | 0.40 | Prices announced in the July 22, 2025 post (audio input reduced by 40% compared with the preview). ### Benchmarks The preview post contains a benchmark table as an image, not included here. ### Key takeaways For classification, translation and routing at scale: the lowest price in the 2.5 range with a 1-million-token window.
Links
Tags : LLM, Google, Gemini, Flash-Lite, Faible coût