FULL STORY
Is Opus 5.5 Secretly Nerfed? Community Fights Back With Benchmarks
After users accused Anthropic of quietly nerfing Claude Opus 5.5, BridgeBench's NerfBench retests found no degradation, and an open-source LiveNerf project now tracks daily benchmark runs to keep watch.
2026-09-26 ~ 2026-09-30 · 2 episodes · 4 posts
Episode 1 · NerfBench retests find no stealth nerfing of Claude Opus 5.5 (2026-09-26, 2 posts)
After users accused Anthropic of stealth-nerfing Claude Opus 5.5, BridgeBench's NerfBench project compared the model's current performance against its launch-day scores. Initial retests found no evidence of degradation, with the model scoring 99.2.
- NerfBench to settle claims that Claude Opus 5.5 was quietly nerfed with day-one vs retest scores — CurieuxExplorer · 2026-09-26
- NerfBench's First Results Find No Nerf: Claude Opus 5.5 Dips Just 0.8% vs Launch — alejandroll10 · 2026-09-28
Episode 2 · LiveNerf: Open-Source Tool Tracks Whether Claude Opus 5.5 Is Being Secretly Nerfed (2026-09-28, 2 posts)
In response to community suspicions that Claude Opus 5.5 has been secretly nerfed, a developer launched LiveNerf, an open-source project that daily reruns independent benchmarks like GPQA and SWE-bench to track the model's performance over time.
- LiveNerf repo tracks Opus 5.5 daily benchmarks to detect if Anthropic nerfs the model — TheOnlyVibemaster · 2026-09-28
- LiveNerf: open-source project tracks day-by-day whether Opus 5.5 gets nerfed — TheOnlyVibemaster · 2026-09-30