Google has officially released stable versions of Gemini 1.5 Flash and 1.5 Pro models, along with a series of API updates and optimizations to the Google AI Studio platform. These updates aim to help developers build and deploy AI applications at a higher efficiency and cost-effectiveness. A notable feature is the significant improvement in rate limits for Gemini 1.5 Flash, now supporting up to 1000 requests per minute (RPM) without a daily request limit. This change is a positive response from Google to developers' needs, aiming to reduce latency and costs in high-volume tasks. While the rate limits for 1.5 Pro remain unchanged, Google encourages developers with higher demands or suggestions to contact them. Starting from June 17th, Gemini 1.5 Flash will introduce model tuning capabilities, allowing developers to customize models for production environments to achieve better performance. This tuning feature will be available in Google AI Studio and Gemini API, and there will be no charge for current tuning jobs or additional token-based billing for using tuned models. To facilitate developers in unlocking higher API rate limits, they can now set up billing accounts in Google AI Studio. For pricing details of Gemini 1.5 models, developers can refer to the Google AI pricing page. If any issues arise during the billing setup process, developers can seek help on the developer forum. For enterprise-level users with specific demands, these models can also be accessed through Vertex AI, Google's enterprise AI platform. Lastly, Google has introduced the JSON Schema feature, allowing developers to specify the desired JSON schema for model responses. This feature opens up new possibilities for use cases that require models to adhere to specific output constraints, such as predefined structures or limited text outputs.