Report
Astra reportedly tops SWE-sweep but finds less than 10% of real-world bugs tested
A post says OpenAI and Anthropic models doubled the number of bugs found while halving costs on SWE-sweep.
TLDR
A post about the updated SWE-sweep leaderboard says OpenAI and Anthropic models doubled the number of bugs found while halving costs. It calls Astra the top model, but says it found less than 10% of all real-world bugs tested. Another poster hopes SWE-sweep will help launch proactive agents that reduce software bugs and vulnerabilities.
Combined views
7.6K
4 Sources, first seen ago
