Google has announced significant updates to its Gemini AI models that introduce improved performance, lower costs, and broader developer access. Google is showing that it is committed to lowering the barrier to entry by offering cheaper, faster, and more efficient models.
Contents
Google has just unveiled two new, production-ready models: Gemini-1.5-Pro-002 and Gemini-1.5-Flash-002 . Both offer significant quality improvements across a variety of tasks, including an impressive 20% increase in math benchmark scores . These models not only handle math better, but also offer twice the speed of results and three times lower latency than previous versions.
Perhaps most importantly for developers building AI apps, Google has significantly reduced the cost of using Gemini 1.5 Pro—by more than 50% for queries and responses under 128K tokens. What’s more, throughput limits have been raised, meaning that you can now do a lot more in less time, opening the door to building more complex apps in less time.
Google is making these changes to make it easier for developers to build intelligent applications. With better understanding of long contexts and improved visual capabilities, the new models are ready to be deployed in even more advanced scenarios. In addition, changes to default filter settings give more control over the configuration of models, which is key to tailoring AI for specific use cases.
“This is a significant step forward in building faster, smarter, and cheaper AI solutions,” said one Google engineer. “We made these changes so developers can build innovative products without worrying about cost or performance.”
While it’s not Gemini 2 yet, the latest version is a big upgrade over the experimental models that were tested before. It allows developers to build apps faster, cheaper, and with greater precision . This opens up new possibilities, especially for startups and smaller companies that may have previously faced financial barriers to developing AI-based solutions.
The new Gemini models can process longer contexts, which is crucial for solutions in the field of analysis of legal documents or financial reports - areas that have huge potential in Poland. Thanks to this, for example, applications supporting lawyers in the analysis of contracts can work faster and with greater precision.