Connect with us

Google

Google drops Gemini 3 Flash, makes it the default AI model

Gemini 3 Flash is based on the Gemini 3 model with a goal to: do more, cost less, and show up everywhere.

Gemini 3 Flash logo on black background
Image: Google

Just a heads up, if you buy something through our links, we may get a small share of the sale. It’s one of the ways we keep the lights on here. Click here for more.

Google drops Gemini 3 Flash, a new AI model that’s faster, cheaper, and very clearly aimed at stealing a little thunder from OpenAI. 

Gemini 3 Flash is based on the Gemini 3 model launched just last month, but it arrives with a clear mission: do more, cost less, and show up everywhere. 

Google is making it the default model in the Gemini app and in AI-powered Search globally, quietly replacing Gemini 2.5 Flash, as nothing dramatic happened. 

Users who want maximum brainpower for math or coding can still switch to the Pro version, but Flash is now the main character.

Performance-wise, this thing punches well above its weight. 

On Humanity’s Last Exam (yes, that’s really what it’s called), Gemini 3 Flash scored 33.7%, nearly tripling the 11% score of its predecessor and landing just shy of Gemini 3 Pro’s 37.5%. 

It even edges past GPT-5.2 in some areas. On the MMMU-Pro benchmark, which tests multimodal reasoning, Gemini 3 Flash straight-up won with an 81.2% score.

Google says Flash shines at understanding what users mean, not just what they type. 

You can upload a pickleball video for tips, doodle a sketch for the AI to guess, toss in an audio clip, and get analysis, or even take a quiz. 

It’s also better at visual answers, mixing images and tables, and can help spin up app prototypes right inside the Gemini app.

Developers get access via Vertex AI, Gemini Enterprise, and a preview API, plus Antigravity, Google’s new coding tool. Big-name companies like JetBrains and Figma are already using it.

At $0.50 per million input tokens and $3.00 per million output tokens, it’s slightly pricier than Gemini 2.5 Flash. 

However, Google claims it’s three times faster than Gemini 2.5 Pro and uses 30% fewer tokens for “thinking” tasks, meaning your bill might actually shrink.

Behind the scenes, the AI rivalry is heating up. Google says it’s now processing over one trillion tokens per day

Meanwhile, Sam Altman reportedly sent a “Code Red” memo after ChatGPT traffic dipped, prompting OpenAI’s own rapid-fire releases.

Google, for its part, is playing it cool. 

According to Tulsee Doshi, this nonstop model leapfrogging is just the industry doing what it does best: pushing the frontier, and occasionally shoving each other while doing it.

Follow us on Flipboard, Google News, or Apple News

Ronil is a Computer Engineer by education and a consumer technology writer by choice. Over the course of his professional career, his work has appeared in reputable publications like MakeUseOf, TechJunkie, GreenBot, and many more. When not working, you’ll find him at the gym breaking a new PR.

Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Limited time TikTok advertising offer details.

More in Google