Overview
Pandas 3.0, released on January 21, 2026, adopts string dtype by default, enables Copy-on-Write and modernizes Series handling.
Main Features
String dtype by default
String columns now use StringDtype by default instead of object, with better memory management and faster operations.
python
import pandas as pd
df = pd.DataFrame({'name': ['Alice', 'Bob', None]})
print(df['name'].dtype) # string (no longer object)
print(df['name'][2]) # <NA> (no longer None)Copy-on-Write by default
Copy-on-Write (CoW) is now enabled by default, eliminating implicit copies and SettingWithCopyWarning.
python
import pandas as pd
df = pd.DataFrame({'a': [1, 2, 3]})
df2 = df[['a']] # view, no copy
df2['a'] = 10 # automatic copy (CoW)
print(df['a'].tolist()) # [1, 2, 3] (unchanged)
print(df2['a'].tolist()) # [10, 10, 10]Series modernization
Series benefit from optimized operations and better integration with Pandas' nullable type system.