RIROBOIND / GLOBAL
← Back to Global Robotics News / Editorial detail
Global Robotics News / EDITORIAL DETAIL

Google’s New Video Model Tops Two Leaderboards and Cuts Prices by 30%

Google’s new video model reportedly leads two public leaderboards while cutting generation prices by about 30%. The article reviews its capabilities, benchmarks, pricing, and implications for video-generation workflows.

From the Temple of Penang
QbitAI is public.

Google, who are you going to kill?

This morning, Google released a new generation of video models.Gemini Omni 1.2 FlashTake it easy.Video from ArenaandTop of the YouTube video doubles list。

图片
图片

What's more, the price is cheaper than the previous generation.30%!

It's just a matter of making a 15-second 720p video.It's $1.05!。 

Who is he who can endure?

Even more activists have taken the lead in playing, bringing out a very soft golden-haired girl.

△Oh, oh, how lovely this little look is!

Friends of video lovers, we have new toys to fold!

For 15 seconds, for 7 bucks, you can change the video like a chat.

This time, Google has really played a new role in AI video generation.

The Gemini Omni 1.2 Flash, which generates the first video, takes a breath away from the previous generation's 10 seconds.15 seconds。

And it doesn't have to be fixed to create a whole video.

We can freely choose the length between 3 seconds and 15 seconds as needed.

图片△From Google Blog

At the same time, prices have also plunged.

Now 720p video per second, as long as it'sIt's $0.07 (about $0.47)It's cheaper than the last generation.30%。

If you want to repeatedly try to create a pointer word, you can create a draft with 360p or 720p, and you'll be happy to enlarge it to 1080p or even 4K.

This way, you don't have to adjust every little bit of detail to earn more money for a whole high-definition video.

图片

And I think the most interesting thing about this upgrade is that it's going to beMulti-round dialogue editing。

The worst thing I've ever seen in an AI video is that it's not easy to generate a satisfactory image, with a biased background, lights, or character clothing that doesn't fit.

If you want to change, you have to re-draw the card.

Right now,Google has also turned the video editing into a conversation.

For example, in the official presentation, all you need to do is follow the instructions and the background can be turned into an oak forest, a white horse, a hot dog market, etc.

The background changed several times throughout, but the two characters' movements, armor reflection and lens composition all kept the same.

There is no need to push back, and there is no need to re-switch characters.

This time, even videos can be discussed with AI.

图片

In addition, the new edition is reinforced.Ability to follow referencesFor the specific effects, you can refer to this test dress demonstration on Google Blog:

What a mess!

So where is this new toy going to play?

At the moment,Gemini API and Google Cloud are officially openIt's a great way to get to know each other.Google AI StudioI'll try it and see how it works out.👀~

Everyone is crazy, and even Frisia is hot.

The new model has been unveiled, and users who are happy to see it are sure to have a lot of active links.

No, it was the two old acquaintances who first came on the scene.

Netizens@chetasluaWith the Gemini Omni Flash 1.2, a hot-blooded version of Frissa from the Battle of the Bulge was released.

And they met face to face in the wilderness of the rock, and the pieces of the rock crumbled, and the gas field collided.

The Golden Light Purple turned up in the lineup, and finally, even the Genghis Khans were organized.

On the other hand, netizens@DesignArenaIn the meantime, the test site was moved directly to the ancient Roman battlefield:

Similarly, the FPV lens instruction, the previous generation Gemini Omni 1.1 Flash, although capable of pulling the lens from underground space gradually out of the battlefield, is more like a continuous retreat and uplift, with some structural changes in motion.

But with Omni 1.2 Flash, it's not the same.

The footage travels through dark stone passages and layered gates, flying high above the outer walls of the bullring, ultimately closing in on the whole bullring and the city of Rome:

It's very smooth.

图片

However, DesignArena also pointed out that Omni 1.2 Flash actually took some time to complete this set of mirrors.

The last generation is time consuming.53 secondsCosts$1.01The new version is time consuming.One minute and 28 seconds.Costs$1.04。

It took more than 35 seconds to replace a more complex and coherent operating mirror.

At the same time, Chetaslua also commented that Omni 1.2 Flash is only a small update to Flash 1.1.

图片

So the new version does have room for improvement.

But the truth has been revealed.

It's a good idea to take a look at the new model and see what it can do.

Welcome to the commentary area and share your feelings of grief with us!

图片

One More Thing

Last but not least, Google hasn't forgotten to provide benefits for everyone.

Now you can go directly to Google Vids and experience the creation of 15-second videos.You don't need to make extra money.

Friends who want to play multiple editing rounds can also go to Flow and touch on new features.

Don't worry about it, Gina!

The link is as follows:
[1]
[2]
[3]