Align mlx::core::min op nan propagation with NumPy

[jhavukainen]

Proposed changes

Follow-up to #2339, extending the same treatment to the min op so NaNs get propagated.

Adds the similar NaN returns to non-integral op types and the specialized treatment for complex types. Extends the reduce nanpropagation tests to also test the min case. Adds the benchmarks for min.

Checklist

Put an x in the boxes that apply.

• [ x ] I have read the CONTRIBUTING document
• [ x ] I have run pre-commit run --all-files to format my code / installed pre-commit prior to committing changes
• [ x ] I have added tests that prove my fix is effective or that my feature works
• [ x ] I have updated the necessary documentation (if needed)

[awni]

Awesome, thanks!

[awni]

The test failure isn't from this diff. We have a flaky quantization test that NaN propagation exposes. But we'll need to fix that in a separate PR.