Andy Halliday

Now, I have a workspace account, an individual workspace account, so I can't test that question.

But if there's a free user out there who wants to download anti-gravity, which is a very cool thing, you can use the view markdown files and so on, but also do agentic coding for you.

Just check and see if you can get access to the top model or whether they basically throttle you as a free user back to, oh, poor you, Gemini 3, which is plenty for you.

Yeah, I have something here.

So let me just...

Also add that there is a new Arc AGI.

So they have an Arc AGI 3 that has been announced.

Yeah.

Just in time.

Just in time.

Let's go back in time, right?

Arc AGI 1 was published in 2019 before GPT-3 came out.

So ArcAGI-1 didn't last very long.

The models progressed so swiftly in 2022, 2023, that it was surpassing the ability of that benchmark to really reasonably differentiate models.

But it's still used out there.

I mean, you can see the ArcAGI-1 leaderboard here.

Then they came out with Arc AGI 2, which is tougher, more difficult visual puzzles and reasoning things that make it not possible.

I shouldn't say not possible, but less likely that the training can be tuned to beat that benchmark.

It requires, you know, real reasoning processing.

Now, RKGI-3 is an interactive reasoning benchmark designed to measure an AI agent's ability to generalize in novel, unseen environments.