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