Livenerf: Has Opus 5.5 been nerfed yet?

Livenerf is a long-running benchmark designed to detect whether AI models degrade in performance after their public launch, addressing widespread reports that Anthropic "nerfs" Claude models post-release. The project started tracking Claude Opus 5.5 on its launch day (September 22, 2026) using deterministic testing methods with frozen prompts, pinned software versions, and statistical analysis to measure any performance drift over 30 days. As of September 29, the benchmark has collected 6 of 10 baseline days without issues, with full results expected after day 20.