After trailing OpenAI for a while, Google finally unleashedGemini, its new powerhouse AI model, last week to revolutionize various aspects of the company’s operations. Gemini is Google’s big move to assert dominance in the AI game. ThePixel 8 Pro now boasts Gemini Nano, supercharging theSummarize feature on Android’s Recorder app, while the mid-tier Gemini Pro is already live on Bard, aligning Google’s AI chatbot with the free version of ChatGPT. Now, developers have their hands on Gemini Pro, ready to experiment with fresh use cases.
Gemini Pro hasrolled outin public preview mode through Vertex AI, Google’s unified AI platform, thanks to the new Gemini Pro API. The API won’t cost you a dime for now, though you’re bound by certain limits. However, Gemini Pro only supports 38 languages and regions at the moment, and throws in a few features like chat, filtering, function calling, embeddings, semantic retrieval, and custom knowledge grounding.
The multimodal AI is kicking off with a 32k context window for text, but future versions are gearing up to crank it up even more. Right now, the model only accepts text input and spits out text-based output, but the Gemini Pro Vision platform welcomes images into the mix.
If you’re diving into Gemini Pro on Vertex AI, input will cost you $0.00025 per 1K characters and $0.0025 per image, while output is billing you at $0.0005 per 1K characters. But Gemini Pro is feeling generous and letting Vertex AI customers test the waters for free until early next year.
AI Studio has a limit of 60 requests per minute, which is a whopping 20 times more than the other freebies out there, as per Google. The company adds that it’s perfect for most of your app development needs.
When you’re done, just click the “get code” button, and you can whisk your work away to your favorite IDE, including JavaScript, Python, cURL, and more. You can also take your pick from the quickstart templates available or toss it straight into Vertex AI.
After Gemini launched a few days ago, itgot its fair share of side-eyefor not being able to handle images, despite proudly sporting the “multimodal” label. Image processing steps in to fix that glitch.
If you’re itching to dive in, the search giant haslaunchedGoogle AI Studio, a free web-based toolkit you’re able to play with right in your browser, provided you’re signed in to your Google account. It allows you to create prompts in a jiffy and snag an API key for app development. Google AI Studio can snatch text and images straight from your Google Drive.
you may hop back and forth between AI Studio and Vertex AI, giving you the option to fine-tune and tweak Gemini, shaping it into the perfect match for your business needs while keeping a tight grip on your data.
