Tag
Xiaomi released MiMo-V2.5-Pro, a coding AI scoring 73.7 on SWE-Bench Pro (near Claude Opus 4.6's 77.1) at 40-60% lower token cost than US frontier models.
Kimi K2.6 open-source model surpasses Opus 4.6 on SWE-Bench, supporting 12+ hour autonomous coding sessions with 4,000+ tool calls.
OpenAI announces it will no longer report SWE-bench Verified scores, citing two critical issues: 59.4% of failed problems have flawed test cases that reject correct solutions, and frontier models have seen benchmark problems during training, making improvements reflect training data exposure rather than genuine capability gains.
Anthropic's updated Claude 3.5 Sonnet achieves a new state-of-the-art 49% on the SWE-bench Verified benchmark, demonstrating significant capabilities in autonomous software engineering tasks.