GPT 6 Sol worse than GPT 5.6 Sol on DeepSWE

On the DeepSWE benchmark, GPT 6 Sol achieved a score of 68.8%, which is lower than GPT 5.6 Sol's score of 72.7%. This indicates that GPT 6 Sol performed worse than its predecessor, GPT 5.6 Sol, on this specific evaluation. The comparison highlights a decrease in performance for the newer model in this context.