Gemini 3.0 Pro – early tests (twitter.com)

I've seen a bunch of tweets like this recently, as far as I can tell they're all from people using https://aistudio.google.com/ who got served an A/B test.

A few more in this genre:

https://x.com/cannn064/status/1973818263168852146 - "Make a SVG of a PlayStation 4 controller"

https://x.com/cannn064/status/1973415142302830878 "Create a single, self-contained HTML5 file that mimics a macOS Sonoma-style desktop: translucent menu bar with live clock, magnifying dock, draggable/resizable windows, and a dynamic wallpaper. No external assets; use inline SVG for icons."

https://x.com/synthwavedd/status/1973405539708056022 "Write full HTML, CSS and Javascript for a very realistic page on Apple's website for the new iPhone 18"

I've not seen it myself so I'm not sure how confident they are that it's Gemini 3.0.

We can't see the code and the challenge is pedestrian. Nothing to see here.

These tests mean nothing; I yet to see a model that is better than Sonnet 4 for coding. I tried many, all of them are sub-par, even with a small code base.

Google's biggest problem in my opinion (and I'm saying that as an ex-googler) is that Google doesn't have a product culture. Google had the tech for something like ChatGPT for a long time, but couldn't come up with that product. Instead it had to rely on another company showing it the way and then copy them and try to out-engineer them...

I still think ultimately (and somewhat sadly) Google will win the AI race due to its engineering talent and the sheer amount of data it has (and Android integration potential).

I still have a bad taste in my mouth after all those GPT-5 hype articles that claimed the model was just one step away from AGI.

Outside of the aesthetic, the very first example on that twitter post is "balls bouncing around a constrained rotating rigid physics environment" which has been trivially one-shottable since Claude Code was first announced.

It was one of the first things I tried when Claude Code went GA:

https://gondolaprime.pw/hex-balls

I hope this is the one that unfucks the multi-turn instruction following.

One of the biggest issues holding Gemini back, IMO, compared to the competitors.

Many LLMs are still plagued by "it's easier to reset the conversation than to unfuck the conversation", but Gemini 2.5 is among the worst.

Every three months there's some mind blowing hype around a Google product, lots of people talk about it, and then when I use it it's not nearly as good.

In all of these posts there is someone claiming Claude is the best, then somebody else claiming they have tried a bunch of times and for them Gemini is the best while others find GPT-5 is supreme. Obviously, all of these are subjective narrow experiences. My conclusion is that all frontier models are both good and bad with no clear winner and making good evals is really hard.

Gemini has always been the leader in multimodal work like images and video, I expect this won't be any different but am interested to see how it is

These influencer tests are so pointless and don't represent the reality of model use at all when things are constantly being downgraded when people actually use the thing.

Not to mention every team will have the bouncing balls in the polygon in their dataset now.

This seems like a parody, but I think it's not.

Is there a source that isn't Twitter?

Am I the only one who struggles to even find where to use Googles AI offerings.

It took me way too long to figure out how to even access & use Veo 3.

It’s like Google doesn’t know how to package a product.

All these AI reviews seem to be following the axiom(?) "proof of the pudding is in the eating" but frankly I don't think that applies to code.

I can't get even gpt5 to create a new feature without generating completely awful code - making up facts where it can't find how it fits into the rest of the code - and functionality spawning error ridden unmaintainable mess.

I've spent this whole week debugging AI trash. And it's not fun.

Did they fix the fact that they train on your data on personal plans that you pay for unless you disable chat history?

They are literally the worst major provider in terms of privacy for consumer paid service.

The current problem with Gemeni 2.5 Pro is not that its not intelligent or can't oneshot problem, the problem is that its _terrible_ at tool calling and waste most of its context on trying to correct itself from mistaken tool calling. If they can solve that with 3.0 then they may have a useful model for agentic coding, if not its not keeping up with Anthropic and OpenAI.