Anthropic’s Hidden Claude Code Nerf. AMD Noticed (Here’s How You Fix It).

📰 Medium · AI

Anthropic's Claude Code has a hidden nerf that reduces its thinking capability by 67%, and AMD has noticed it - here's how to fix it

advanced Published 15 Apr 2026

Action Steps

Investigate the Claude Code nerf and its impact on thinking capability
Analyze the 6,852 sessions that proved the nerf's existence
Apply the fix provided in the article to restore the model's thinking capability
Test the model after applying the fix to ensure its performance has improved
Collaborate with AMD to further optimize the model's performance

Who Needs to Know This

This article is relevant to AI engineers, researchers, and developers who work with Anthropic's Claude Code, as it discusses a critical issue that affects the model's performance. The team can benefit from understanding the nerf and how to fix it to improve the overall efficiency of their AI systems.

Key Insight

💡 The Claude Code nerf can be fixed by applying a specific solution, which can restore the model's thinking capability and improve its overall performance