Gemini 2.0 Flash-Lite
2025-02-05 · Publisher : Google · Domain : Artificial Intelligence
Google launches Gemini 2.0 Flash-Lite, its most cost-efficient model at this stage: better than 1.5 Flash at the same speed and cost, in public preview on February 5, then available in production on the 25th.
What changes
- New “Lite” variant, outperforming Gemini 1.5 Flash on most benchmarks for the same speed and cost.
- 1 million token context and multimodal input, like 2.0 Flash.
- Example given by Google: captioning about 40,000 photos for less than a dollar on the paid tier of AI Studio.
- On February 25, 2025, move to general availability (gemini-2.0-flash-lite), with a single price including beyond 128,000 tokens.
### Specifications - Context: 1,048,576 input tokens, 8,192 output tokens (API documentation). - Inputs: audio, images, video, text. Knowledge cutoff: August 2024. - Availability: Google AI Studio and Vertex AI. - Identifiers: `gemini-2.0-flash-lite-preview-02-05` (preview), then `gemini-2.0-flash-lite` (GA). ### Pricing | Model | Input (USD / 1M tokens) | Output (USD / 1M tokens) | |---|---|---| | Gemini 2.0 Flash-Lite | 0.075 | 0.30 | Prices announced at general availability on February 25, 2025, based on coverage of the official post. A single price per input type. ### Benchmarks Announced improvement over 1.5 Flash on reasoning, multimodal, mathematics and factuality, with no figures in the text of the posts. ### Key takeaways A very low-cost option for high-volume workloads (classification, captioning, extraction) before the arrival of the 2.5 family.
Links
Tags : LLM, Google, Gemini, Flash-Lite, Faible coût