I'm super excited,

every time I get a new model to test, it's like Christmas.

(upbeat music)

For every new model release we get at Lovable,

we run a series of benchmarks and internal evaluations,

and running what we call the hard tasks on GPT-5.5

is when we saw that there was

a pretty big step in capabilities.

One thing that we see across projects

is that 5.5 is a lot better at planning,

which means that for large features,

our users are much more likely to succeed in one shot

rather than having to ask multiple times for iterations.

With GPT-5.5, we've seen a 31% increase

in intent understanding during planning,

and 22% fewer amnesia, instances of amnesia,

where we measure if models are kind

of forgetting information from their context.

And that's really important

as you go deeper into a large session

where you're working on complex features.

I think the magic of Lovable is that our users

don't need to think really about anything

besides their goal,

and let them build whatever they want

without them needing to think about the code at all.

(upbeat music continues)