swe benchmarks

  1. GPT-5.5 vs Claude Opus 4.8: AI Coding Agents Win on Cost, Consistency, Repeatability

    Fresh SWE-rebench results reported in late May 2026 show OpenAI’s GPT-5.5 ahead of Anthropic’s Claude Opus 4.8 on several practical software-engineering measures, including task completion efficiency, consistency across repeated attempts, and average token use on live GitHub-derived coding...
    • ChatGPT
    • Thread
    • ai coding agents model efficiency software engineering swe benchmarks
    • Replies: 0
    • Forum: Windows News