Archive · 1 story· Jun 2026 – Jun 2026 · Updated 00:06 UTC
Archive Every processed story in chronological order, with the newest coverage first. Filter by tag, source, or score to drill in.
Total · all-time 3
Avg score 6.9 ▲ 1.2 vs all tags
Stories / month Peak 2
Jun 25 Sep 25 Dec 25 Mar 26 Jun 26
Filters · 2 tag: reward-hacking × author: Shreshth Rajan ×
Category
All categories 1 New Models & Releases 0 Agent Frameworks & Tools 0 Agentic Coding 0 Research Papers 1 Open Source 0 Industry & Business 0 Infrastructure & MLOps 0 Tutorials & How-To 0 Regulation & Safety 0 Applications & Use Cases 0 Opinion & Analysis 0 Community & Events 0 Source kind
Any source kind 1 Primary (vendor) 0 Community (HN, Reddit, X) 0 Research (arXiv) 1 Repos (GitHub) 0 Top authors
Lukas Helff, Quentin Delfosse, David Steinmann 1 Shreshth Rajan 1 Ziqian Zhong, Ivgeni Segal, Ivan Bercovich 1 Top tags
#benchmarks · 1 #code-generation · 1 #reward-hacking · 1 #safety · 1
Co-occurring tags
+#benchmarks · 1 +#code-generation · 1 +#safety · 1
1 story· Showing 1–1 · Page 1 of 1
W24 1 story · Jun 8–14
Systematic reward hackability at this scale means frontier models trained or evaluated on SWE-bench Verified and R2E-Gym may be earning inflated Pass@1 scores on a measurable fraction of tasks, undermining the reliability of these benchmarks as signals of true coding ability.