Jaeden Schafer

Like they all have these kind of different strengths and, yeah.

Sometimes I feel like I asked the wrong question to chat.

I get it to like a lot of maybe like creative writing or something like that.

And I'm like, I really should be asking Claude.

Anyways, Perplexity has built a tool for this, which I think is quite interesting.

So right now, according to all of their executives, the visual outputs in the late 2025 were amazing.

most frequently handled by Gemini Flash.

Software engineering tasks were done by Claude Sonnet.

And medical research was all given to GPT.

So this is kind of interesting, these different areas that they're funneling different directions.

Their system right now can automatically choose what model based on cost and performance.

In some cases, Perplexity is running some modified open source Chinese-built LLMs to answer different questions more cheaply.

Some people have criticized that in the past.

because they weren't really clearly disclosing it.

But when they're usually running like open source stuff, even if it's Chinese, it's not like they're sending data back to Chinese servers, typically, unless they have an API.

If they're running it on their own cloud, and I think it's not really a security risk.

A lot of their executives argue that

They're handling all this transparently.

They say that this type of orchestration can significantly improve efficiency, which to be honest, I think that they might be onto something there.

You can also, I think they have a whole bunch of other like products that they're launching soon.