I don't understand this assertion, but maybe I'm missing something? Google included a SWE-bench score of 63.8% in their announcement for Gemini 2.5 Pro: https:

I don't understand this assertion, but maybe I'm missing something? Google included a SWE-bench score of 63.8% in their announcement for Gemini 2.5 Pro: https://blog.google/technology/google-deepmind/gemini-model-thinking-updates-march-2025/#gemini-2-5-pro https://blog.google/technology/google-deepmind/gemini-model-...