How can we batch requests?

[thisismygitrepo]

First check

• I added a descriptive title to this issue.
• I used the GitHub search to look for a similar issue and didn't find it.
• I searched the Marvin documentation for this feature.

Describe the current behavior

It is cheaper if we batch in regular llm frameworks, just wondering if this is doable in case of marvin.

Describe the proposed behavior

as above.

Example Use

No response

Additional context

No response

[zzstoatzz]

hi @thisismygitrepo - thanks for the issue!

do you have a DX in mind?

[thisismygitrepo]

Hi. Not really, but judging by superficial understanding of the innerworkings of marvin, it is piggybacking off pydandtic_ai models, meaning, problem solved if pydnatic_ai solves it?