Google is introducing Gemini 3 Flash, making it a standard feature in Gemini software

Google today released its fastest and cheapest version of Gemini 3 Flash, based on Gemini 3 released last monthlooking to steal OpenAI’s thunder. The company is also making this a standard feature in the Gemini software and AI search engine.

The new version of Flash will arrive six months after Google announced it Gemini 2.5 Flash model, very interesting. On the benchmark, the Gemini 3 Flash model surpasses the one established by its limitations and matches the performance of other models, such as Gemini 3 Pro and GPT 5.2, in some ways.

For example, it achieved 33.7% without using a tool Humanities Final Exam, which is designed to measure expertise in various domains. In comparison, Gemini 3 Pro scored 37.5%, Gemini 2.5 Flash scored 11%, and the newly released GPT-5.2 scored 34.5%.

On the multimodality and reasoning benchmark MMMU-Pro, the new model surpassed all competitors with a score of 81.2%.

Consumer release

Google is making Gemini 3 Flash the default version of Gemini software worldwide, replacing Gemini 2.5 Flash. Users can select the Pro version from the math and math selectors.

The company says the new model is better at identifying content in multiple formats and giving you a response based on that. For example, you can upload your short pickleball video and ask for tips; you can try to draw and have a model imagine what you are drawing; or you can upload a recording to analyze or create questions.

The company also said that the model better understands user queries and can create visual solutions with content such as images and tables.

You can also use a new model making app prototypes in the Gemini program using instructions.

Gemini 3 Pro is now available for anyone in the US to search and many people in the US can also find the Nano Banana Pro image by searching.

Access to companies and developers

Google noted that companies like JetBrains, Figma, Cursor, Harvey, and Latitude are already using the Gemini 3 Flash model, which is available through Vertex AI and Gemini Enterprise.

For developers, the company is making the model available in preview mode through API and Antigravity, A new Google tool was released last month.

The company said that the Gemini 3 Pro will score 78% on the SWE-bench benchmark, bested only by the GPT-5.2. It added that the model is good for video analysis, data extraction, and visual Q&A, and because of its speed, it is suitable for fast and repetitive motion.

Sample prices are $0.50 for 1 million input tokens and $3.00 for 1 million output tokens. This is slightly more expensive than $0.30 for 1 million input tokens and $2.50 for 1 million output tokens for Gemini Flash 2.5. But Google claims that the new version outperforms the Gemini 2.5 Pro version at three times faster. And, for logic tasks, it uses 30% fewer tokens on average than 2.5 Pro. This means that in general, you can save on the number of tokens for other services.

If you look, for example, even if the prices are higher than the top of this table, Flash is the cheapest from the point of view of prices and results.

Since releasing Gemini 3, Google has processed more than 1 trillion tokens per day on its API, amid its aggressive release and operational battle with OpenAI.

Earlier this month, Sam Altman reportedly posted inside “Code Red” memo to the OpenAI team later ChatGPT traffic has decreased as Google’s market share among consumers rose. Post this, OpenAI has it GPT-5.2 release and a new image format. OpenAI also boasted about the growth of its business and said that the number of ChatGPT messages is there 8x growth from November 2024.

Although Google did not directly focus on competition with OpenAI, it said that the release of new models will make it difficult for all companies to work.

“What’s happening in all these companies is like all these brands continue to be amazing, to challenge each other, to push the boundaries. And I think what’s also amazing is that companies are bringing out these brands,” Doshi said.

“We’re also bringing in new indicators and new ways to evaluate these species. And that’s encouraging as well.”