Editor’s note: The post has been updated to reflect that 1.5 Flash tuning support has been delayed a few weeks and will not launch on June 17.
Building on the momentum from Google I/O, we're announcing important updates to the Gemini API and Google AI Studio, including:
We’re incredibly excited to see what you build with these new models and are committed to building towards a world class developer experience. You can get started with Gemini 1.5 Flash and 1.5 Pro free of charge in Google AI Studio.
Gemini 1.5 Flash was purpose-built as our fastest, most cost-efficient model yet for high volume tasks, at scale, to address developers’ feedback asking for lower latency and cost. Today, we are increasing the rate limit for 1.5 Flash to 1000 requests per minute (RPM) and removing the request per day limit. The 1.5 Pro rate limit will not be changed at this time, but if you need even higher limits to scale or have feedback, please reach out to us.
Customizing models can help you reach the performance threshold needed to take AI models into production. To support that, we will also be rolling out tuning support for Gemini 1.5 Flash in the coming weeks. Tuning will be supported in both Google AI Studio and the Gemini API directly. Currently, tuning jobs are free of charge, and using a tuned model does not incur any additional per-token costs. You can learn more about tuning in the Gemini API docs.
In addition to the free tier, starting today, developers can unlock higher API rate limits by turning on a billing account in Google AI Studio.
You can learn more about the Gemini 1.5 model pricing on ai.google.dev/pricing. If you run into any issues setting up billing, please let us know on our developer forum. For developers looking to scale with enterprise-grade features, the same models are available via Vertex AI, our enterprise-ready AI platform.
We launched JSON mode in the Gemini API and Google AI Studio earlier this year to give you more control over model output. Starting today, you can specify the desired JSON schema for the model to respond with, which unlocks many new use cases where you need the model to conform to certain output constraints like following a predefined structure or only outputting specific text. You can read more about JSON schema mode in the Gemini API docs.
To give developers more flexibility in AI Studio, you can now choose your preferred UI mode (light vs dark) or use your system defaults in the settings pane. We also rolled out our first set of mobile improvements for Google AI Studio to allow you to quickly test multi modal prompts on-the-go.
As we continue to improve our developer experience, please share your feedback on our Developer Forum. Happy building!
