DeepSWE adds GPT-5.6 models to benchmark

DeepSWE added GPT-5.6 models to their coding benchmark, according to a Reddit post. The chart shows performance comparisons against existing agents.