The default model has been established... Google ignites the AI ​​race with the launch of Gemini 3 Flash.

 

Google Gemini 3 Flash

The default model has been established... Google ignites the AI ​​race with the launch of Gemini 3 Flash.

Google has announced the launch of its new Gemini 3 Flash model, which is fast and low-cost, in a move aimed at strengthening its position in the artificial intelligence race and directly competing with OpenAI.

The company decided to adopt the new model as the default model within the Gemini application and put artificial intelligence into the search engine.
Gemini 3 Flash comes just about one month after the launch of the Gemini 3 family, and six months after the release of Gemini 2.5 Flash, with significant improvements in performance and speed.

According to benchmark tests, the new model clearly outperformed its predecessor, achieving results close to advanced models such as Gemini 3 Pro and GPT-5.2 in a number of indicators, according to a report published by TechCrunch and reviewed by Al Arabiya Business.

In Humanity's Last Exam, which measures knowledge across multiple domains, the Gemini 3 Flash scored 33.7% without the use of tools, compared to 37.5% for the Gemini 3 Pro and only 11% for the Gemini 2.5 Flash, while the GPT-5.2 scored 34.5%.
In the MMMU-Pro test for reasoning and multimedia, the new model outperformed all competitors with a score of 81.2%.

Wide access for users

Google explained that it has begun adopting Gemini 3 Flash as the default model for Gemini users worldwide, while retaining the Gemini 3 Pro option for advanced math and programming questions.
The new model features improved capabilities in understanding multimedia content, such as analyzing short videos, interpreting hand-drawn graphics, and analyzing audio files, in addition to generating more visually rich answers through tables and images.
The model also allows for the creation of application prototypes directly from within Gemini using only text commands.

Google has confirmed that companies such as JetBrains, Figma, Cursor, and Harvey have already begun using Gemini 3 Flash across their Vertex AI and Gemini Enterprise platforms.
The model has also become available to developers in a trial version via the API, in addition to the new Antigravity coding tool.

As for pricing, the model costs $0.50 per million input codes and $3 per million output codes, which is slightly higher than Gemini 2.5 Flash, but Google asserts that it is three times faster and consumes 30% less code in complex reasoning tasks, which may reduce the overall cost in the long run.

A fierce race with "OpenAI"

This launch comes amid fierce competition with OpenAI, as Google revealed that it currently processes more than a trillion codes per day through its APIs.
In contrast, OpenAI has recently intensified its releases with GPT-5.2 and new image generation models.
Although Google did not comment directly on the competition, it confirmed that the rapid launch of advanced models is driving all companies to continuous innovation, indicating the intensifying race to lead the future of artificial intelligence.