Google today released its fast and affordable Gemini 3 Flash model, building on the Gemini 3 foundation from last month. This move aims to compete directly with OpenAI. The company is also making this new model the default in the Gemini app and within the AI features in search.
The arrival of Gemini 3 Flash comes six months after Google announced the Gemini 2.5 Flash model, and it offers substantial improvements. In benchmark tests, the Gemini 3 Flash model significantly outperforms its predecessor and matches the performance of other leading models like Gemini 3 Pro and GPT-5.2 in certain areas.
For example, on the Humanity’s Last Exam benchmark, which tests broad expertise, Gemini 3 Flash scored 33.7% without using external tools. 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 outscored all competitors with a score of 81.2%.
Google is rolling out Gemini 3 Flash as the default model in the Gemini app globally, replacing Gemini 2.5 Flash. Users can still select the Pro model from the menu for tasks like math and coding.
The company states the new model excels at identifying multimodal content and providing answers based on it. You can upload a short pickleball video for tips, draw a sketch for the model to guess, or upload an audio recording for analysis or quiz generation. Google also said the model better understands user intent and can generate more visual answers, incorporating elements like images and tables.
You can also use the new model to create app prototypes in the Gemini app using simple prompts.
The Gemini 3 Pro model is now available to everyone in the U.S. for search, and more people in the U.S. can access the Nano Banana Pro image model in search as well.
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 a preview through the API and in Antigravity, Google’s new coding tool released last month.
Google said the Gemini 3 Pro scores 78% on the SWE-bench verified coding benchmark, a score only outperformed by GPT-5.2. It added that the model is ideal for video analysis, data extraction, and visual Q&A, and its speed makes it suited for quick, repeatable workflows.
Model pricing is set at $0.50 per 1 million input tokens and $3.00 per 1 million output tokens. This is slightly more expensive than the Gemini Flash 2.5 pricing of $0.30 per 1 million input tokens and $2.50 per 1 million output tokens. However, Google claims the new model outperforms the Gemini 2.5 Pro model while being three times faster. For thinking tasks, it uses 30% fewer tokens on average than 2.5 Pro, which could lead to overall token savings for certain tasks.
Company executives position Flash as a workhorse model, noting its lower cost makes it suitable for bulk tasks for many companies.
Since releasing Gemini 3, Google has processed over 1 trillion tokens per day on its API, amid a fierce release and performance competition with OpenAI.
Earlier this month, reports indicated OpenAI’s CEO sent an internal “Code Red” memo after ChatGPT’s traffic dipped as Google’s consumer market share rose. Following that, OpenAI released GPT-5.2 and a new image generation model. OpenAI also highlighted its growing enterprise use and stated that ChatGPT message volume has grown eightfold since November 2024.
While Google did not directly address the competition with OpenAI, it commented that the release of new models is challenging all companies to be active and push the frontier, and that the introduction of new benchmarks is encouraging further advancement.