Lengthy, but there might not be a better alternative without fixing the data.

CREATE TABLE SampleData(
  col1 varchar(10),
  col2 varchar(10),
  col3 int
);
INSERT INTO SampleData
VALUES
  ('A', 'B', 10),
  ('B', 'A', 5 ),
  ('C', 'D', 8 );

SELECT CASE WHEN col1 <= col2 THEN col1 ELSE col2 END AS col1,
   CASE WHEN col1 <= col2 THEN col2 ELSE col1 END AS col2,
   SUM( col3) AS col3
FROM SampleData
GROUP BY CASE WHEN col1 <= col2 THEN col1 ELSE col2 END,
   CASE WHEN col1 <= col2 THEN col2 ELSE col1 END;

GO
DROP TABLE SampleData