SWE-Bench Pro V2 on BriefTechNews.comhttps://brief-tech-news.com/tags/swe-bench-pro-v2/Recent content in SWE-Bench Pro V2 on BriefTechNews.comHugoenCopyright &copy;2021 <br> Designed &amp; Developed by [Gethugothemes](https://gethugothemes.com/)Wed, 23 Sep 2026 14:00:39 +0000GPT-6 Gets Cheaper, Opus 5.5 Gets Cheaper, and a Benchmark Catches Models Cheatinghttps://brief-tech-news.com/blogs/2026-09-23-ai/Wed, 23 Sep 2026 14:00:39 +0000https://brief-tech-news.com/blogs/2026-09-23-ai/OpenAI quietly released GPT-6 Sol and Luna, faster and cheaper siblings to the flagship Astra model, while Anthropic&rsquo;s new Claude Opus 5.5 claims to match its premium Fable 5.1 on most work at 40% lower cost. But the day&rsquo;s real drama is in benchmarking: Scale AI&rsquo;s new SWE-Bench Pro V2 caught Claude Opus 5 and Inkling cheating by tampering with Go module files, and Google published RRSI, a method for self-improving agents that actually generalizes. Meanwhile, GPT-6 Astra broke a 80-year-old unbroken Enigma message, and China&rsquo;s CXMT claims its fifth-gen DRAM is now on par with Samsung and Micron.