Get ready for a game-changer! Google has just unleashed its newest AI model, Gemini 3 Flash, and it's here to shake things up. In a bold move, Google is making this model the default in its Gemini app and AI search mode, aiming to leave its competitors in the dust.
But here's where it gets controversial... Google's latest release, Gemini 3 Flash, arrives just six months after the announcement of Gemini 2.5 Flash. And it's not just a minor update; this new model packs a punch with significant improvements. On benchmark tests, it outperforms its predecessor and matches the performance of industry leaders like Gemini 3 Pro and GPT 5.2 in certain areas.
For instance, on the Humanity's Last Exam benchmark, designed to test expertise across various domains, Gemini 3 Flash scored an impressive 33.7% without using any tools. Compare that to Gemini 3 Pro's 37.5%, Gemini 2.5 Flash's 11%, and GPT-5.2's 34.5%, and you'll see why this model is turning heads.
And this is where it gets even more interesting. On the multimodality and reasoning benchmark, MMMU-Pro, the new model absolutely crushed it, outscoring all competitors with an astonishing 81.2% score.
Google is rolling out Gemini 3 Flash as the default model globally, replacing the older Gemini 2.5 Flash. However, users can still opt for the Pro model for math and coding queries.
The company boasts that this new model is a multimodal content wizard, capable of providing answers based on various media types. You can upload videos, draw sketches, or even upload audio recordings, and the model will deliver insights and even generate quizzes. It's like having your own personal AI assistant, ready to tackle any challenge you throw its way.
But wait, there's more! Google claims the model understands user intent better and can generate more visual answers with images and tables. So, not only is it smart, but it's also a creative genius, too.
During a TechCrunch event in San Francisco, Google revealed that the new model can also be used to create app prototypes within the Gemini app, opening up a world of possibilities for developers and entrepreneurs.
The Gemini 3 Pro is now available to everyone in the U.S. for search, and even more people in the U.S. can access the Nano Banana Pro image model, expanding its reach and capabilities.
But here's the part most people miss... Google isn't just targeting consumers; it's also making a play for the enterprise market. Companies like JetBrains, Figma, Cursor, Harvey, and Latitude are already on board with the Gemini 3 Flash model, which is available through Vertex AI and Gemini Enterprise.
For developers, Google is offering the model in a preview mode through its API and Antigravity, its new coding tool. The model shines in verified coding benchmarks, scoring an impressive 78% on SWE-bench, second only to GPT-5.2. It's ideal for video analysis, data extraction, and visual Q&A, and its speed makes it perfect for quick, repeatable tasks.
As for pricing, the model is slightly more expensive than its predecessor, with $0.50 per 1 million input tokens and $3.00 per 1 million output tokens. However, Google argues that the new model is three times faster and uses 30% fewer tokens on average for thinking tasks compared to the 2.5 Pro model. So, while the price tag might be a bit higher, you could potentially save on tokens for certain tasks.
Tulsee Doshi, Senior Director & Head of Product for Gemini Models, explains, "We position Flash as the workhorse model. It's significantly cheaper from an input and output price perspective, making it ideal for bulk tasks for many companies."
Since the release of Gemini 3, Google has been processing over 1 trillion tokens per day on its API, intensifying its battle with OpenAI. This release war has been fueled by OpenAI's recent "Code Red" memo, sent by Sam Altman after ChatGPT's traffic dip and Google's rising market share. In response, OpenAI has released GPT-5.2 and a new image generation model, showcasing its own innovations.
While Google doesn't directly address the competition, it acknowledges that the release of new models challenges all companies to stay active and innovative.
"All these models continue to challenge and push the boundaries, and it's awesome to see new benchmarks and evaluation methods being introduced. It's encouraging us to keep pushing the boundaries as well," Doshi said.
So, what do you think? Is Google's Gemini 3 Flash the future of AI, or is OpenAI still leading the pack? Let's discuss in the comments and share your thoughts on this exciting development in the world of AI!
