{"slug": "ai-coding-agents-still-need-human-cleanup-on-nearly-three-out-of-four-tasks-in-a", "title": "AI coding agents still need human cleanup on nearly three out of four tasks in a real-world test", "summary": "Alibaba's new Co-Design Bench, which tests AI coding agents on real user requests that had already tripped them up in practice, found that even the top performer, GPT-6 Astra, fully solves only 26.4 percent of tasks. The same models score above 99 percent on older benchmarks, meaning human cleanup is still needed on nearly three out of four real-world tasks.", "body_md": "Alibaba's new coding-agent benchmark, Co-Design Bench, tests AI agents on real user requests that already tripped them up in practice. Even GPT-6 Astra, the top performer, fully solves only 26.4 percent of tasks. On older benchmarks, the same models score above 99 percent.\n\nThe article [AI coding agents still need human cleanup on nearly three out of four tasks in a real-world test](https://the-decoder.com/ai-coding-agents-still-need-human-cleanup-on-nearly-three-out-of-four-tasks-in-a-real-world-test/) appeared first on [The Decoder](https://the-decoder.com).", "url": "https://wpnews.pro/news/ai-coding-agents-still-need-human-cleanup-on-nearly-three-out-of-four-tasks-in-a", "canonical_source": "https://the-decoder.com/ai-coding-agents-still-need-human-cleanup-on-nearly-three-out-of-four-tasks-in-a-real-world-test/", "published_at": "2026-11-08 09:50:59+00:00", "updated_at": "2026-11-09 01:28:14.447176+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-agents", "ai-tools", "large-language-models"], "entities": ["Alibaba", "Co-Design Bench", "GPT-6 Astra", "The Decoder"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/ai-coding-agents-still-need-human-cleanup-on-nearly-three-out-of-four-tasks-in-a", "markdown": "https://wpnews.pro/news/ai-coding-agents-still-need-human-cleanup-on-nearly-three-out-of-four-tasks-in-a.md", "text": "https://wpnews.pro/news/ai-coding-agents-still-need-human-cleanup-on-nearly-three-out-of-four-tasks-in-a.txt", "jsonld": "https://wpnews.pro/news/ai-coding-agents-still-need-human-cleanup-on-nearly-three-out-of-four-tasks-in-a.jsonld"}}