ImportError: cannot import name 'XGBClassifier' from 'xgboost'

TL;DR: your xgboost is ancient; the sklearn-style XGBClassifier was added later. Upgrade xgboost and the import works.

The error

ImportError: cannot import name 'XGBClassifier' from 'xgboost'

Fix it

1. Upgrade xgboost:

   pip install -U xgboost

   Expected: installs a modern release.

2. Verify:

   python -c "from xgboost import XGBClassifier; print('ok')"

   Expected: prints ok.

When this applies

Use this when code does from xgboost import XGBClassifier on a very old xgboost.

When it does not apply

If xgboost is already modern and the import fails, the install is corrupt; reinstall. If the error is the native-library load failure instead, that is a separate missing-.so problem.

Compatibility

xgboost 0.7+. XGBClassifier has existed for years; only truly ancient installs lack it.

Root cause

Early xgboost exposed only the low-level xgb.train API. The sklearn-compatible wrappers came later, so installs from that era have no XGBClassifier name to import.

Edge cases

Upgrading across many major versions can change default parameters; validate model metrics after the jump. If a project pins the old xgboost deliberately, the sklearn API simply is not available there.

GitHub issue dmlc/xgboost#2847 reports ImportError: cannot import name 'XGBClassifier' from 'xgboost' with an old xgboost installed. The verified fix: the scikit-learn wrapper API (XGBClassifier, XGBRegressor) was added after that old release, so upgrading xgboost to a modern version provides the names. Full thread: https://github.com/dmlc/xgboost/issues/2847