Adding finalization to objects created for every async for loop is likely to have an adverse effect on performance.
Have you benchmarked this?

Running https://github.com/python/pyperformance/blob/main/pyperformance/data-files/benchmarks/bm_async_generators/run_benchmark.py shows Benchmark hidden because not significant (1): async_generators. My benchmarking setup isn't very accurate though.

I plan to land this later today if there are no further concerns.

🤖 New build scheduled with the buildbot fleet by @kumaraditya303 for commit 5e21e60 🤖

If you want to schedule another build, you need to add the 🔨 test-with-buildbots label again.