Gemini 1.5 Pro Updated, 1.5 Flash Launched, 2 New Gemma Models Unveiled

Hey there, tech enthusiasts! If you're looking to dive into the world of AI, grab yourself an API key from Google AI Studio and get ready to explore the Gemini API Cookbook. Last December, we rolled out Gemini 1.0, making the Gemini API accessible in Google AI Studio and Vertex AI for all you enterprise folks. Fast forward to February, and we introduced Gemini 1.5 Pro, the first model with a whopping 1 million token context window. And just last month, we spiced things up with native audio understanding, system instructions, JSON mode, and more.

But wait, there's more! Alongside Gemini, we launched Gemma, our family of lightweight open models, crafted from the same tech and research behind Gemini. Since February, the 2B and 7B models have been downloaded like hotcakes, and we threw in some cool variants last month like CodeGemma and RecurrentGemma for all you developers and researchers out there.

Now, buckle up because Gemini and Gemma are about to get even better. We're rolling out updates to Gemini 1.5 Pro, introducing 1.5 Flash, and adding some new developer features and two fresh Gemma models. Let's dive into the details.

Gemini 1.5 Pro Improvements and a New 1.5 Flash Model

• Gemini 1.5 Pro: We've been tinkering with this model, making quality improvements across the board—from translation to coding and reasoning. These updates are live now, so you can tackle even bigger and more complex tasks.
• Gemini 1.5 Flash: This smaller model is perfect for those quick, high-frequency tasks where speed is key.
• Availability: Both models are up for grabs in over 200 countries and territories in preview mode, and they'll be fully available come June.
• Natively Multimodal with Long Context: Both 1.5 Pro and 1.5 Flash come with our 1 million token context window, letting you mix text, images, audio, and video as inputs. Want to get your hands on 1.5 Pro with a 2 million token context window? Just hop on the waitlist in Google AI Studio or Vertex AI if you're a Google Cloud customer.

New Developer Features and Pricing Options for the Gemini API

• New Developer Features: We've been listening to your feedback, and we're rolling out video frame extraction and parallel function calling, which lets you return multiple function calls at once. And guess what? In June, we're adding context caching to Gemini 1.5 Pro, so you only need to send parts of your prompt, including those hefty files, once. This should make the long context even more useful and wallet-friendly.
• Pricing: We're keeping the Gemini API free in eligible regions through Google AI Studio, and we're boosting the rate limits with our new pay-as-you-go service. Check out the latest prices for Google AI Studio and Vertex AI.

Additions to the Gemma Family

• PaliGemma: Our first vision-language open model is here today, optimized for image captioning, visual Q&A, and other image labeling tasks. It joins our other pre-trained Gemma variants, CodeGemma and RecurrentGemma.
• Gemma 2: Coming in June, our next-gen Gemma is built for top-notch performance at sizes that developers love. You've been asking for a bigger model that's still easy to use, and the new Gemma 27B model fits the bill. It outperforms some models that are more than twice its size and runs smoothly on GPUs or a single TPU host in Vertex AI.

The First Gemini API Developer Competition

Get ready, because we're launching our first-ever Gemini API Developer Competition! We want to see your most creative and impactful apps built with Gemini models. Submit your projects by August 12, and you might just win the grand prize—a custom electric DeLorean!

Want more stories from Google? Sign up for our newsletter and get the latest updates straight to your inbox.

Ready to build with Gemini? Give 1.5 Pro and 1.5 Flash a whirl in Google AI Studio today.