Gemini 3 Flash is significantly faster and more efficient than 2.5 Pro. Watch below as 3 Flash generates complex graphics, 3D models, and a web app before the previous generation even finishes processing⚡

Great work, Sundar. Flash is amazing! 👌

Bonus: We published this UI on AIStudio for anyone to run these comparisons! ⚡⚡⚡

faster models = agents that can actually do things in real-time

Why are you comparing to 2.5 Pro? No one uses that anymore. How does it compare to 3 Pro?

This matters. Faster models unlock entirely new workflows, not just nicer demos.

But 2.5 pro did not ignore my instructions to give the whole updated files when asking for changes, and it did not miss files that need updating...

⚡️⚡️⚡️

Efficiency beats elegance at scale. Always has.

I know thinking budget is largely configurable through API, but I hope the default tendencies of 3-Flash is not to overthink much. One of my fav things about 2.0-flash (or lite) was how quick it could to generate reasonable objects.

It's wild how fast Google's making their smartest AI models cheaper. Can't wait for even faster and cheaper ones down the road!

also wild that Flash is now good enough to handle things that used to be "Pro-only" territory : SWE‑bench scores, complex tool use, rich visualizations - without melting your token budget.

Speed is crucial. The quickest models dominate distribution, and Flash seems designed for practical use in the real world, not just for show.

Speed like that changes how you build iteration becomes instant, and ideas go from concept to working output without the usual waiting.

Google is cooking

AI speed upgrades now feel like watching a toddler outrun a marathoner.

Ok Flash Thinking is dope as hell (Pro fan here) but what’s with the “Would you like me to…?” follow up question at the end of each answer? That was GPT-5 failure. Gemini is too cool for that! Please fix. Thank you!

If speed and efficiency keep improving this fast, does model choice start becoming more about latency and cost than raw capability?

Impressive speed and efficiency upgrade! Great progress.

Gemini 3 Flash is the new default for vibe coding 👀

Retail still struggling to open Excel while 3 Flash is building 3D worlds.

Speed is a feature, but not the feature. This is about latency disappearing.

Crushing it, boss! When 3 flash images?

Impresionante avance. La eficiencia y velocidad de Gemini 3 Flash son notables. Sin duda, esto marca una nueva etapa en el desarrollo de gráficos y modelos 3D.

it mean's u guys also reduces token usages..Great work🔥🔥🔥🔥🔥

really weird comparison

Flash didn’t wait for permission. It shipped.

Wow this is nice, need to try

How is the cost comparison between 2.5 pro Vs 3 flash?

Is @grok better than Gemini 3?

That side-by-side demo is brutal – 3 Flash finished a full web app while 2.5 Pro was still loading images! Google's speed game is unmatched right now. Congrats on the launch!

The speed difference is impressive, but I'm curious about how it handles more nuanced creative tasks where thoughtfulness matters more than raw speed.