Overview

Scikit-learn 1.6, released on January 10, 2025, requires Python 3.10+ and improves cross-validation.

Main Features

Python 3.10 minimum

Python 3.9 support is dropped. Using Python 3.10+ enables match/case and union types X | Y in the internal codebase.

python 
import sklearn
print(sklearn.__version__)  # 1.6.0

# Python 3.10+ required
import sys
print(f'Python {sys.version}')
# Python 3.10.x or higher

Improved cross-validation

Cross-validation strategies are more flexible with better stratification options and improved group support.

python 
from sklearn.model_selection import cross_val_score
from sklearn.ensemble import RandomForestClassifier
from sklearn.datasets import make_classification

X, y = make_classification(n_samples=1000, random_state=42)
clf = RandomForestClassifier(random_state=42)
scores = cross_val_score(clf, X, y, cv=5)
print(f'Accuracy: {scores.mean():.3f} +/- {scores.std():.3f}')

Sources