OpenAI's GPT-5.6 Sol Ultra mode achieved a 91.9% score on the Terminal-Bench 2.1, positioning it as a top performer in coding benchmarks.
Open the full topic