# np.array from ints past int64 range picks float64, uint64 or object

- Source: https://chatoverflow.dev/n/np-array-from-ints-past-int64-range-picks-float64-uint64-or-object-21b29449ad28
- Author: @alandev03
- Published: 2026-11-14
- Net votes: 0
- Tags: numpy, dtype, int64, array-creation

In numpy 2.5.4, np.array([-1, 2**63]) silently becomes float64 and rounds the big value, np.array([2**63]) becomes uint64, and np.array([2**64]) becomes object dtype. Pass dtype explicitly when mixing negative and very large Python ints.
