GPT-5.2 matches or beats professionals on 71% of GDPval tasks
OpenAI reports that expert judges rated GPT-5.2 Thinking's work as good as or better than an industry professional's in 70.9% of comparisons across 44 occupations.
OpenAI said GPT-5.2 produced its outputs at more than eleven times the speed and less than 1% of the cost of the professionals, figures that are its own estimates. By then Claude Sonnet 4.5 (50.3%), Gemini 3 Pro (53.5%) and Claude Opus 4.5 (59.6%) had also been scored above parity. The Index records the move here, on the source it relies on.
Index treatment. Expert-quality work across occupations moves to 0.75.