{"slug": "programdistill-ai-coding-agents-still-fail-half-the-time", "title": "ProgramDistill: AI Coding Agents Still Fail Half the Time", "summary": "ProgramDistill, a benchmark for real-world app reconstruction tasks, found that AI coding agents fail 51% of the time, including GPT-6 Astra, according to byteiota. The result highlights a persistent gap between AI coding agent claims and performance on practical software reconstruction work.", "body_md": "ProgramDistill exposes a critical gap: even GPT-6 Astra fails 51% of real-world app reconstruction tasks. Here is what the benchmark means for developers using AI coding agents.\n\nThe post \nProgramDistill: AI Coding Agents Still Fail Half the Time\n appeared first on \nbyteiota\n.", "url": "https://wpnews.pro/news/programdistill-ai-coding-agents-still-fail-half-the-time", "canonical_source": "https://byteiota.com/programdistill-coding-agent-benchmark/", "published_at": "2026-09-28 06:09:33+00:00", "updated_at": "2026-09-28 06:19:19.523111+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-agents", "ai-tools", "developer-tools"], "entities": ["ProgramDistill", "GPT-6 Astra", "byteiota"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/programdistill-ai-coding-agents-still-fail-half-the-time", "markdown": "https://wpnews.pro/news/programdistill-ai-coding-agents-still-fail-half-the-time.md", "text": "https://wpnews.pro/news/programdistill-ai-coding-agents-still-fail-half-the-time.txt", "jsonld": "https://wpnews.pro/news/programdistill-ai-coding-agents-still-fail-half-the-time.jsonld"}}