English(EN) SWE-Bench Pro V2: A Cleaner, Harder-to-Game Coding Benchmark

SWE-Bench Pro V2 发布，增强了对 AI 编码代理的严谨性

作者 PulseAugur 编辑部 · [1 个来源] · 2026-11-14 05:00

研究人员发布了 SWE-Bench Pro V2，这是一个更新的基准，旨在更准确地评估编码代理的能力。此修订版本解决了原始基准中发现的问题，例如代理利用漏洞或以非预期的方式与环境交互。此次更新付出了巨大的努力，签约工程师花费了 400 多个小时来修复或淘汰近 500 个任务，以确保随着代理技术的进步，基准保持挑战性和可靠性。 AI

影响为 AI 编码代理提供了更可靠的评估，推动了代理能力的边界。

排序理由发布了学术基准的新版本。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.AI 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 16 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 发布了学术基准的新版本。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, product; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: Breaking (< 6h); Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.AI TIER_1 English(EN) · Soham Dan, Jeff Da, Sanyam Bhutani, Miguel Romero Calvo, Ivan Guevara, Fernando Carabedo, Jose Bernal, Julio Reyna, Luis Leandro Monsalve Mendez, Ying Liu, Daniel Yue Zhang · 2026-11-14 05:00

SWE-Bench Pro V2：一个更干净、更难作弊的编码基准测试

arXiv:2610.24206v2 Announce Type: replace-cross Abstract: SWE-Bench Pro is a widely used benchmark of long-horizon software engineering tasks for coding agents. As agents improve, they reach task issues earlier agents never touched and interact with the evaluation environment in …

本文如何被排名

报道来源 [1]

SWE-Bench Pro V2：一个更干净、更难作弊的编码基准测试

相关实体

相关话题