Home› 🧠 Models ›Why we no longer evaluate SWE-bench Veri… OpenAI 🧠 Models 🕐 Feb 23, 2026 👁 81 ⏱ 1 min read Why we no longer evaluate SWE-bench Verified Annons Annons SWE-bench Verified is increasingly contaminated and mismeasures frontier coding progress. Our analysis shows flawed tests and training leakage. We recommend SWE-bench Pro. Source: OpenAI — Published February 23, 2026 at 11:00 UTC — Category: Models 🔗 Read full article on OpenAI → ⚡ Popular — 81 views Share: 𝕏 Twitter in LinkedIn 💬 WhatsApp 🔗 Copy Link More in Models The Decoder18h ago Anthropic pushes ahead with pre-Thanksgiving IPO despite tech selloff, accepts lower valuation The Decoder19h ago AI auditors catch most agent misbehavior but flood reviewers with false alarms, new study finds The Decoder19h ago Thinking Machines' Inkling 2 Small drops 316 F-bombs across benchmark runs The Decoder20h ago The gap between open and closed AI models in cyber capabilities keeps shrinking The Decoder23h ago State propaganda in some Chinese AI models is only "a thin layer" that fine-tuning can strip away The Decoder1d ago Nine AI agents tried autonomous research on open questions from US national labs and none passed the experts' bar Annons Annons