I went to check and Opus 5.5 wasn't benchmarked on DeepSWE1.1, I can't keep track of which are in vogue at the moment but maybe it's thought to be saturated at this point, as they only have older GPT models as well

Artificial analysis has benched it on deepswe