DeepSWE blows up AI coding leaderboard, crowns GPT-5.5, + ClaudeOpus loophole(venturebeat.com)
venturebeat.com
DeepSWE blows up AI coding leaderboard, crowns GPT-5.5, + ClaudeOpus loophole
https://venturebeat.com/technology/deepswe-blows-up-the-ai-coding-leaderboard-crowns-gpt-5-5-and-finds-claude-opus-exploiting-a-benchmark-loophole
1 comments
Hmm... If this holds up it could explain a lot of things.