Meanwhile, Google released a new AI Character builder and Gemini 3 Flash that makes some of that stuff pretty easy. This new model increases performance and efficiency across both Gemini app. This newest model comes just six months after the release of Gemini 2.5 Flash. Not only will it replace its predecessor as the default option for users in the United States of America and across the globe, Gemini 3 Flash is intended to provide better quality output at a reduced cost per operation. This strategic focus helps it emerge as a top contender among all the other powerful AI models emerging at warp speed.
The new model provides some meaningful upgrades, with full tool use scoring a remarkable 33.7% (no tool use) on the Humanity’s Last Exam benchmark. It performed best in the multimodality and reasoning benchmark, MMMU-Pro. With a stellar score of 81.2%, it beat out all rival models in the area on the environment. The benefits of this lofty performance increase are substantial. Gemini 3 Flash requires 30% fewer tokens for the types of thinking tasks compared to the full Gemini 2.5 Pro model.
Pricing and Availability
The new pricing structure for Gemini 3 Flash accommodates its more advanced capabilities. It’s currently $0.50 / 1 million input tokens, $3.00 / 1 million output tokens. SmartChat w/ Gemini 2.5 Flash charges $0.30 per million input tokens, $2.50 per million output tokens. Yet, with its more affordable pricing, Google is actively pushing Gemini 3 Flash as the preferred option for large-scale jobs.
Tulsee Doshi, senior director and head of Product for Gemini Models, highlighted the importance of the model to business needs.
“We really position flash as more of your workhorse model. So if you look at, for example, even the input and output prices at the top of this table, Flash is just a much cheaper offering from an input and output price perspective. And so it actually allows for, for many companies, bulk tasks.” – Tulsee Doshi
Additionally, businesses can develop application prototypes directly within prompts in the Gemini app, emphasizing the model’s adaptability and real-world application.
Performance Enhancements
Google says that Gemini 3 Flash beats the old Gemini 2.5 Pro model on performance while running three times faster. This major speed advantage is a big deal for users who need super quick responses in applications ranging from search capabilities, code completion and creative work.
In just four days since Gemini 3’s release, Google said it’s experienced incredible user engagement with the model. They’ve been doing over 1 trillion tokens a day on their API – wow! With Gemini 3 Pro officially released, we’re excited to introduce it to users around the U.S. for search. This trailblazing launch extends powerful new options to anybody looking to leverage Google’s cutting-edge generative AI capabilities.
Doshi took us on a journey through the development of these models, discussing how this evolution is affecting the industry environment.
“Just about what’s happening across the industry is like all of these models are continuing to be awesome, challenge each other, push the frontier.” – Tulsee Doshi
The Gemini app is loaded with stellar new functionality! Users can still select the Pro model from the model picker for math and coding questions, providing a range to better serve their personal needs.
Industry Impact and Future Directions
The arrival of Gemini 3 Flash marks a different turn that highlights Google’s continued and expanded offerings. Taking this step would be a savvy play to remain competitive in an increasingly crowded market. New AI models keep coming out. In turn, Google is committed to establishing new standards and creating ways of testing their models.
“We’re also introducing new benchmarks and new ways of evaluating these models. And so that’s also encouraging us.” – Tulsee Doshi
This commitment indicates that Google is not just focused on performance but is actively seeking ways to enhance user experience and adapt to changing industry standards.