That's the kind of transparency we need! That DeepSWE benchmark puts it in frontier territory: https://artificialanalysis.ai/agents/coding-agents?coding-ag...