https://venturebeat.com/technology/deepswe-blows-up-the-ai-coding-leaderboard-crowns-gpt-5-5-and-finds-claude-opus-exploiting-a-benchmark-loopholeVentureBeaten2026-05-26T22:32:44.818+00:00DeepSWE blows up the AI coding leaderboard, crowns GPT-5.5, and finds Claude Opus exploiting a benchmark loopholehttps://images.ctfassets.net/jdtwqhzvc2n1/4kUVtxUVBjivIKlO68RxPf/88918e60aed6f6c50fb031ea81e52f8f/deepswe-card.jpg?w=300&q=30