本文介绍了 Gemini 2.0 模型家族的最新更新，现已通过谷歌 AI 工作室和顶点 AI 提供。文章重点介绍了三个关键变体：Gemini 2.0 Flash（具有增强功能的通用版本）、Flash-Lite（适用于大规模文本输出的成本高效选项）以及 Pro（针对编码和复杂任务优化的实验性更新）。此外，最近推出的 Gemini 2.0 Flash Thinking Experimental 被强调为一项重要补充，它在响应前提供推理能力。与 Gemini 1.5 相比，这些模型在性能上有显著提升，支持多模态输入，简化了定价结构并降低了成本。开发者可以利用谷歌 AI 工作室和顶点 AI 等工具，将这些模型无缝集成到他们的工作流程中。性能基准和定价细节通过图表展示，体现了技术深度和实际优势。
Gemini 2.0 Family expands for developers
Today we announced exciting updates that make Gemini 2.0 available to more developers and production use. The following models are now available in the Gemini API via Google AI Studio and in Vertex AI:
- Gemini 2.0 Flash is now generally available, with higher rate limits, stronger performance, and simplified pricing.
- Gemini 2.0 Flash-Lite, a new variant that is our most cost-efficient model yet, is now available in public preview.
- Gemini 2.0 Pro, an experimental update to our best model yet for coding and complex prompts, is now available.
Together with the recently-launched Gemini 2.0 Flash Thinking Experimental, our Flash variant that reasons before answering, these releases make Gemini 2.0 capabilities available to a broad range of use cases and applications.
Model features
Gemini 2.0 Flash offers a comprehensive suite of features, including native tool use, a 1 million token context window, and multimodal input. It currently supports text output, with image and audio output capabilities and the Multimodal Live API planned for general availability in the coming months. Gemini 2.0 Flash-Lite is cost-optimized for large scale text output use cases.
Model performance
The Gemini 2.0 models deliver significant performance improvements over Gemini 1.5 across a range of benchmarks.
As with prior models, Gemini 2.0 Flash defaults to a concise style that makes it easier to use and reduces cost. It can also be prompted to use a more verbose style that produces better results in chat-oriented use cases.
Gemini Pricing
We continue to reduce costs with Gemini 2.0 Flash and 2.0 Flash-Lite. Both have a single price per input type, removing the Gemini 1.5 Flash distinction between short and long context requests. This means the cost of both 2.0 Flash and Flash-Lite can be lower than Gemini 1.5 Flash with mixed-context workloads, despite the performance improvements that both deliver.
Follow these links to learn more about token counting for different modalities, about Gemini Developer API pricing, and about Vertex AI pricing.
You can start building with the latest Gemini models in four lines of code, with an industry leading free tier and rate limits to scale to production. We’re inspired by your progress thus far and can’t wait to see how you will use these latest Gemini models. Happy building!
