OpenAI reports that GPT-5.6 Sol has reduced prompt injection failures by 6 times compared to the best production model from 4 months ago

Odaily Planet Daily News: OpenAI posted on the X platform that training against GPT-Red has significantly improved the robustness of GPT-5.6. OpenAI replayed some of the strongest attacks from GPT-Red for measurement, and none of these related attacks had been seen by the model during training. GPT-5.6 Sol has become OpenAI's most robust model against prompt injection to date, with the number of failures reduced by 6 times compared to the best production model from 4 months ago.