Context

We should run a subset of benchmarks on PRs, so that we can run them multiple times and average out the results to reduce variability. We can then use those results in PYTHON-5513 to get a quick look at potential regressions introduced by a PR. We'll need to update the Python Triage Context to target this new task after we create it.

Definition of done

Create the new task and update the Python Triage Context.

Pitfalls

None